EMMA: Extensible MultiModal Annotation markup language

W3C Working Draft 9 April 2007

This version:
http://www.w3.org/TR/2007/WD-emma-20070409/
Latest version:
http://www.w3.org/TR/emma/
Previous version:
http://www.w3.org/TR/2005/WD-emma-20050916/
Editor in chief:
Michael Johnston, AT&T
Editors:
Paolo Baggia, Loquendo
Jerry Carter, Nuance
Deborah A. Dahl, Independent Consultant
Gerry McCobb, IBM
Dave Raggett, W3C/Volantis

Abstract

The W3C Multimodal Interaction working group aims to develop specifications to enable access to the Web using multimodal interaction. This document is part of a set of specifications for multimodal systems, and provides details of an XML markup language for containing and annotating the interpretation of user input. Examples of interpretation of user input are a transcription into words of a raw signal, for instance derived from speech, pen or keystroke input, a set of attribute/value pairs describing their meaning, or a set of attribute/value pairs describing a gesture. The interpretation of the user's input is expected to be generated by signal interpretation processes, such as speech and ink recognition, semantic interpreters, and other types of processors for use by components that act on the user's inputs such as interaction managers.

Status of this Document

This section describes the status of this document at the time of its publication. Other documents may supersede this document. A list of current W3C publications and the latest revision of this technical report can be found in the W3C technical reports index at http://www.w3.org/TR/.

This is the 9 April 2007 W3C Last Call Working Draft of "EMMA: Extensible MultiModal Annotation markup language". The Last Call period ends 30 April 2007.

Publication as a Working Draft does not imply endorsement by the W3C Membership. This is a draft document and may be updated, replaced or obsoleted by other documents at any time. It is inappropriate to cite this document as other than work in progress.

This specification describes markup for representing interpretations of user input (speech, keystrokes, pen input etc.) together with annotations for confidence scores, timestamps, input medium etc., and forms part of the proposals for the W3C Multimodal Interaction Framework.

This document has been produced as part of the W3C Multimodal Interaction Activity, following the procedures set out for the W3C Process, with the intention of advancing it along the W3C Recommendation track. The authors of this document are members of the W3C Multimodal Interaction Working Group (members only).

The major changes to this document since 16 September 2005 Last Call Working Draft are the addition of the emma:dialog-turn attribute and the emma:signal-size attribute. Also the specification of emma:uninterpreted has been revised so that it applies to recordings, in which case no interpretation was intended. Also emma:interpretation elements with the emma:unintepreted="true" attribute are required to be empty. The values of emma:mode have also been revised. speech has been replaced with voice and dtmf_keypad with dtmf for consistency with the Voice Browser specifications. Also, the definition of emma:mode and emma:medium have been revised so that they can contain a space delimited set of values. This enables direct annotation of the mode and medium on composite multimodal inputs. Also Section 4.3 on annotations in composite derivations has been removed and the relevant content on the determination of annotations on composite inputs has been added into Section 4.1.2 and Section 4.2.10.4.

This Last Call working draft also incorporates numerous editorial revisions and clarifications made in response to detailed feedback from the Internationalization working group, the Web Accessibility Initiative, the Voice Browser working group, and the Smartweb consortium. These changes include marking of normative versus informative material and use of the key words as described in [RFC2119], to clarify the status of statements in the specification. The references have also been updated.

Several of the features in the current draft specification are considered to be at risk of removal due to potential lack of implementations. The relative timestamp attributes are considered to be at risk: emma:time-ref-uri, emma:time-ref-anchor-point, emma:offset-to-start (Section 4.2.10.2). The emma:hook attribute is considered to be at risk (Section 4.2.12, Appendix C). The container element emma:group and associated element emma:group-info are considered to be at risk (Section 3.3.2). The container element emma:sequence is considered to be at risk (Section 3.3.3). The elements and attributes associated with description of endpoints are considered to be at risk (emma:endpoint-info, emma:endpoint (Section 4.1.5), emma:endpoint-role, emma:endpoint-address, emma:port-type, emma:port-num, emma:message-id, emma:service-name, emma:endpoint-pair-ref, emma:endpoint-info-ref (Section 4.2.14)).

This document was produced by a group operating under the 5 February 2004 W3C Patent Policy. W3C maintains a public list of any patent disclosures made in connection with the deliverables of the group; that page also includes instructions for disclosing a patent. An individual who has actual knowledge of a patent which the individual believes contains Essential Claim(s) must disclose the information in accordance with section 6 of the W3C Patent Policy.

Your feedback is welcomed. Please send feedback to the public mailing list: www-multimodal@w3.org (public archives). See W3C mailing list and archive usage guidelines.

Conventions of this Document

All sections in this specification are normative, unless otherwise indicated. The informative parts of this specification are identified by "Informative" labels within sections.

The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this document are to be interpreted as described in [RFC2119].

Table of Contents

1. Introduction

This section is Informative.

This document presents an XML specification for EMMA, an Extensible MultiModal Annotation markup language, responding to the requirements documented in Requirements for EMMA [EMMA Requirements]. This markup language is intended for use by systems that provide semantic interpretations for a variety of inputs, including but not necessarily limited to, speech, natural language text, GUI and ink input.

It is expected that this markup will be used primarily as a standard data interchange format between the components of a multimodal system; in particular, it will normally be automatically generated by interpretation components to represent the semantics of users' inputs, not directly authored by developers.

The language is focused on annotating the interpretation information of single and composed inputs, as opposed to (possibly identical) information that might have been collected over the course of a dialog.

The language provides a set of elements and attributes that are focused on accurately representing annotations on the input interpretations.

An EMMA document can be considered to hold three types of data:

Given the assumptions above about the nature of data represented in an EMMA document, the following general principles apply to the design of EMMA:

The annotations of EMMA should be considered 'normative' in the sense that if an EMMA component produces annotations as described in Section 3 and Section 4, these annotations must be represented using the EMMA syntax. The Multimodal Interaction Working Group may address in later drafts the issues of modularization and profiling, that is: which sets of annotations are to be supported by which classes of EMMA component.

1.1 Uses of EMMA

The general purpose of EMMA is to represent information automatically extracted from a user's input by an interpretation component, where input is to be taken in the general sense of a meaningful user input in any modality supported by the platform. The reader should refer to the sample architecture in W3C Multimodal Interaction Framework [MMI Framework], which shows EMMA conveying content between user input modality components and an interaction manager.

Components that generate EMMA markup:

  1. Speech recognizers
  2. Handwriting recognizers
  3. Natural language understanding engines
  4. Other input media interpreters (e.g. DTMF, pointing, keyboard)
  5. Multimodal integration component

Components that use EMMA include:

  1. Interaction manager
  2. Multimodal integration component

Although not a primary goal of EMMA, a platform may also choose to use this general format as the basis of a general semantic result that is carried along and filled out during each stage of processing. In addition, future systems may also potentially make use of this markup to convey abstract semantic content to be rendered into natural language by a natural language generation component.

1.2 Terminology

anchor point
When referencing an input interval with emma:time-ref-uri, emma:time-ref-anchor-point allows you to specify whether the referenced anchor is the start or end of the interval.
annotation
Information about the interpreted input, for example, timestamps, confidence scores, links to raw input, etc.
composite input
An input formed from several pieces, often in different modes, for example, a combination of speech and pen gesture, such as saying "zoom in here" and circling a region on a map.
confidence
A numerical score describing the degree of certainty in a particular interpretation of user input.
data model
For EMMA, a data model defines a set of constraints on possible interpretations of user input.
derivation
Interpretations of user input are said to be derived from that input, and higher levels interpretations may be derived from lower level ones. EMMA allows you to reference the user input or interpretation a given interpretation was derived from, see semantic interpretation.
dialog
For EMMA, dialog can be considered as a sequence of interactions between the users and the application.
endpoint
In EMMA, this refers to a network location which is the source or recipient of an EMMA document. It should be noted that the usage of the term "endpoint" in this context is different from the way that the term is used in speech processing, where it refers to the end of a speech input.
gestures
In multimodal applications gestures are communicative acts made by the user or application. An example is circling an area on a map to indicate a region of interest. Users may be able to gesture with a pen, keystrokes, hand movements or sound. Gestures often form part of composite input. Application gestures are typically animations and/or sound effects.
grammar
A set of rules that describe a sequence of tokens expected in a given input. These can be used by speech and handwriting recognizers to increase recognition accuracy.
handwriting recognition
The process of converting pen strokes into text.
ink recognition
This includes the recognition of handwriting and pen gestures.
input cost
In EMMA, this refers to a numerical measure indicating the weight or processing cost associated with a user's input or part of their input.
input device
The device proving a particular input, for example, a microphone, a pen, a mouse, a camera, or a keyboard.
input function
In EMMA, this refers to the use a particular input is serving, for example, as part of a recording or transcription, as part of a dialog, or as a means to verify the user's identity.
input medium
Whether the input is acoustic, visual, or tactile, for instance, a spoken utterance is an example of an aural input, a hand gesture as seen by a camera is an example of a visual input, pointing with a mouse or pen is an example of a tactile input.
input mode
This distinguishes a particular means of providing an input within a general input medium, for example, speech, DTMF, ink, key strokes, video, photograph, etc.
input source
This is the device that provided the input, for example a particular microphone or camera. EMMA allows you to identify these with a URI.
input tokens
In EMMA, this refers to a sequence of characters, words or other discrete units of input.
instance data
A representation in XML of an interpretation of user input.
interaction manager
A processor that determines how an application interacts with a user. This can be at multiple levels of abstraction, for example, at a detailed level, determining what prompts to present to the user and what actions to take in response to user input, versus a higher level treatment in terms of goals and tasks for achieving those goals. Interaction managers are typically event driven.
interpretation
In EMMA, an interpretation of user input refers to information derived from the user input that is meaningful to the application.
keystroke input
Input provided by the user pressing on a sequence of keys (buttons), such as a computer keyboard or keypad.
lattice
A set of nodes interconnected with directed arcs such that by following an arc, you can never find yourself back at a node you have already visited (i.e. a directed acyclic graph). Lattices provide a flexible means to represent the results of speech and handwriting recognition, in terms of arcs representing words or character sequences. Different arcs from the same node represent different local hypotheses as to what the user said or wrote.
metadata
Information describing another set of data, for instance, a library catalog card with information on the author, title and location of a book. EMMA is designed to support input processors in providing metadata for interpretations of user input.
multimodal integration
The process of combining inputs from different modes to create an interpretation of composite input. This is also sometimes refered to as multimodal fusion.
multimodal interaction
The means for a user to interact with an application using more than one mode on interaction, for instance, offering the user the choice of speaking or typing, or in some cases, allowing the user to provide a composite input involving multiple modes.
natural language understanding
The process of interpreting text in terms that are useful for an application.
N-best list
An N-best list is a list of the most likely hypotheses for what the user actually said or wrote, where N stands for an integral number such as 5 for the 5 most likely hypotheses.
raw signal
An uninterpreted input, such as an audio waveform captured from a microphone.
semantic interpretation
A normalized representation of the meaning of a user input, for instance, mapping the speech for "San Francisco" into the airport code "SFO".
semantic processor
In EMMA, this refers to systems that can derive interpretations of user input, for instance, mapping the speech for "San Francisco" into the airport code "SFO".
signal interpretation
The process of mapping a discrete or continuous signal into a symbolic representation that can be used by an application, for instance, transforming the audio waveform corresponding to someone saying "2005" into the number 2005.
speech recognition
The process of determining the textual transcription of a piece of speech.
speech synthesis
The process of rendering a piece of text into the corresponding speech, i.e. synthesizing speech from text.
text to speech
The process of rendering a piece of text into the corresponding speech.
time stamp
The time that a particular input or part of an input began or ended.
URI: Uniform Resource Identifier
A URI is a unifying syntax for the expression of names and addresses of objects on the network as used in the World Wide Web (RFC3986). A URI is defined as any legal anyURI primitive as defined in XML Schema Part 2: Datatypes Second Edition Section 3.2.17 [SCHEMA2].
user input
An input provided by a user as opposed to something generated automatically.

2. Structure of EMMA documents

This section is Informative.

As noted above, the main components of an interpreted user input in EMMA are the instance data, an optional data model, and the metadata annotations that may be applied to that input. The realization of these components in EMMA is as follows:

An EMMA interpretation is the primary unit for holding user input as interpreted by an EMMA processor. As will be seen below, multiple interpretations of a single input are possible.

EMMA provides a simple structural syntax for the organization of interpretations and instances, and an annotative syntax to apply the annotation to the input data at different levels.

An outline of the structural syntax and annotations found in EMMA documents is as follows. A fuller definition may be found in the description of individual elements and attributes in Section 3 and Section 4.

From the defined root node emma:emma the structure of an EMMA document consists of a tree of EMMA container elements (emma:one-of, emma:sequence, emma:group) terminating in a number of interpretation elements (emma:interpretation). The emma:interpretation elements serve as wrappers for either application namespace markup describing the interpretation of the users input or an emma:lattice element or emma:literal element . A single emma:interpretation may also appear directly under the root node.

To illustrate this here is an example EMMA document for input to a flight reservation application. In this example there are two speech recognition results and associated semantic representations of the input. The system is uncertain whether the user meant "flights from Boston to Denver"or "flights from Austin to Denver". The annotations to be captured are timestamps and confidence scores for the two inputs.

Example:

<emma:emma version="1.0"
    xmlns:emma="http://www.w3.org/2003/04/emma"
    xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
    xsi:schemaLocation="http://www.w3.org/2003/04/emma
     http://www.w3.org/TR/2007/WD-emma-20070409/emma.xsd"
    xmlns="http://www.example.com/example">
  <emma:one-of id="r1" emma:start="1087995961542" emma:end="1087995963542">
    <emma:interpretation id="int1" emma:confidence="0.75"
    emma:tokens="flights from boston to denver">
      <origin>Boston</origin>
      <destination>Denver</destination>
    </emma:interpretation>

    <emma:interpretation id="int2" emma:confidence="0.68"
    emma:tokens="flights from austin to denver">
      <origin>Austin</origin>
      <destination>Denver</destination>
    </emma:interpretation>
  </emma:one-of>
</emma:emma>

Attributes on the root emma:emma element indicate the version and namespace. The emma:emma element contains an emma:one-of element which contains a disjunctive list of possible interpretations of the input. The actual semantic representation of each interpretation is within the application namespace. In the example here the application specific semantics involves elements origin and destination indicating the origin and destination cities for looking up a flight. The timestamp is the same for both interpretations and it is annotated using values in milliseconds in the emma:start and emma:end attributes on the emma:one-of. The confidence scores and tokens associated with each of the inputs are annotated using the EMMA annotation attributes emma:confidence and emma:tokens on each of the emma:interpretation elements.

2.1 Data model

An EMMA data model expresses the constraints on the structure and content of instance data, for the purposes of validation. As such, the data model may be considered as a particular kind of annotation (although, unlike other EMMA annotations, it is not a feature pertaining to a specific user input at a specific moment in time, it is rather a static and, by its very definition, application-specific structure). Its specification in EMMA is optional.

Since Web applications today use different formats to specify data models, e.g. XML Schema Part 1: Structures Second Edition [XML Schema Structures], XForms 1.0 (Second Edition) [XFORMS], RELAX NG Specification [RELAX-NG], etc., EMMA itself is agnostic to the format of data model used.

Data model definition and reference is defined in Section 4.1.1.

2.2 EMMA namespace prefixes

An EMMA attribute is qualified with the EMMA namespace prefix if the attribute can also be used as an in-line annotation on elements in the application's namespace. Most of the EMMA annotation attributes in Section 4.2 are in this category. An EMMA attribute is not qualified with the EMMA namespace prefix if the attribute only appears on an EMMA element. This rule ensures consistent usage of the attributes across all examples.

3. EMMA structural elements

This section defines elements in the EMMA namespace which provide the structural syntax of EMMA documents.

3.1 Root element: emma:emma

Annotation emma:emma
Definition The root element of an EMMA document.
Children The emma:emma element MUST immediately contain a single emma:interpretation element or EMMA container element: emma:one-of, emma:group, emma:sequence. It MAY also contain an optional single emma:derivation element and an optional single emma:info annotation element. It MAY also contain multiple optional emma:grammar annotation elements, emma:model annotation elements, and emma:endpoint-info annotation elements.
Attributes
  • Required:
    • version: the version of EMMA used for the interpretation(s). Interpretations expressed using this specification MUST use 1.0 for the value.
    • Namespace declaration for EMMA, see below.
  • Optional:
    • any other namespace declarations for application specific namespaces.
Applies to None

The root element of an EMMA document is named emma:emma. It holds a single emma:interpretation or EMMA container element (emma:one-of, emma:sequence, emma:group). It MAY also contain a single emma:derivation element containing earlier stages of the processing of the input (See Section 4.1.2). It MAY also contain an optional single annotation element: emma:info and multiple optional emma:grammar, emma:model, and emma:endpoint-info elements.

It MAY hold attributes for information pertaining to EMMA itself, along with any namespaces which are declared for the entire document, and any other EMMA annotative data. The emma:emma element and other elements and attributes defined in this specification belong to the XML namespace identified by the URI "http://www.w3.org/2003/04/emma". In the examples, the EMMA namespace is generally declared using the attribute xmlns:emma on the root emma:emma element. EMMA processors MUST support the full range of ways of declaring XML namespaces as defined by the Namespaces in XML 1.1 (Second Edition) [XMLNS]. Application markup MAY be declared in an explicit application namespace, or an undefined namespace (equivalent to setting xmlns="").

For example:

<emma:emma version="1.0" xmlns:emma="http://www.w3.org/2003/04/emma">
    ....
</emma:emma>

or


<emma version="1.0" xmlns="http://www.w3.org/2003/04/emma">
    ....
</emma>

3.2 Interpretation element: emma:interpretation

Annotation emma:interpretation
Definition The emma:interpretation element acts as a wrapper for application instance data or lattices.
Children The emma:interpretation element MUST immediately contain either application instance data, or a single emma:lattice element, or a single emma:literal element, or in the case of uninterpreted input or no input emma:interpretation MUST be empty. It MAY also contain multiple optional emma:derived-from elements and an optional single emma:info element.
Attributes
  • Required: Attribute id of type xsd:ID that uniquely identifies the interpretation within the EMMA document.
  • Optional: The annotation attributes: emma:tokens, emma:process, emma:no-input, emma:uninterpreted, emma:lang, emma:signal, emma:signal-size, emma:media-type, emma:confidence, emma:source, emma:start, emma:end, emma:time-ref-uri, emma:time-ref-anchor-point, emma:offset-to-start, emma:duration, emma:medium, emma:mode, emma:function, emma:verbal, emma:cost, emma:grammar-ref, emma:endpoint-info-ref, emma:model-ref, emma:dialog-turn.
Applies to The emma:interpretation element is legal only as a child of emma:emma, emma:group, emma:one-of, emma:sequence, or emma:derivation.

The emma:interpretation element holds a single interpretation represented in application specific markup, or a single emma:lattice element, or a single emma:literal element.

The emma:interpretation element MUST be empty if it is marked with emma:no-input="true" (Section 4.2.3). The emma:interpretation element MUST be empty if it has been annotated with emma:uninterpreted="true" (Section 4.2.4) or emma:function="recording" (Section 4.2.11).

Attributes:

  1. id a REQUIRED xsd:ID value that uniquely identifies the interpretation within the EMMA document.
<emma:emma version="1.0" xmlns:emma="http://www.w3.org/2003/04/emma"
    xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
    xsi:schemaLocation="http://www.w3.org/2003/04/emma
     http://www.w3.org/TR/2007/WD-emma-20070409/emma.xsd"
    xmlns="http://www.example.com/example">
  <emma:interpretation id="r1">
    ...
  </emma:interpretation>
</emma:emma>

3.3 Container elements

3.3.1 emma:one-of element

Annotation emma:one-of
Definition A container element indicating a disjunction among a collection of mutually exclusive interpretations of the input.
Children The emma:one-of element MUST immediately contain a collection of one or more emma:interpretation elements or container elements: emma:one-of, emma:group, emma:sequence . It MAY also contain multiple optional emma:derived-from elements and an optional single emma:info element.
Attributes
  • Required:
    • Attribute id of type xsd:ID
    • The attribute disjunction-type MUST be present if emma:one-of is embedded within emma:one-of. The possible values of disjunction-type are {recognition, understanding, multi-device, and multi-process}.
  • Optional:
    • On a single non-embedded emma:one-of the attribute disjunction-type is optional.
    • The following annotation attributes are optional: emma:tokens, emma:process, emma:lang, emma:signal, emma:signal-size, emma:media-type, emma:confidence, emma:source, emma:start, emma:end, emma:time-ref-uri, emma:time-ref-anchor-point, emma:offset-to-start, emma:duration, emma:medium, emma:mode, emma:function, emma:verbal, emma:cost, emma:grammar-ref, emma:endpoint-info-ref, emma:model-ref, emma:dialog-turn.
Applies to The emma:one-of element MAY only appear as a child of emma:emma, emma:one-of, emma:group, emma:sequence, or emma:derivation.

The emma:one-of element acts as a container for a collection of one or more interpretation (emma:interpretation) or container elements (emma:one-of, emma:group, emma:sequence), and denotes that these are mutually exclusive interpretations.

An N-best list of choices in EMMA MUST be represented as a set of emma:interpretation elements contained within an emma:one-of element. For instance, a series of different recognition results in speech recognition might be represented in this way.

<emma:emma version="1.0"
    xmlns:emma="http://www.w3.org/2003/04/emma"
    xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
    xsi:schemaLocation="http://www.w3.org/2003/04/emma
     http://www.w3.org/TR/2007/WD-emma-20070409/emma.xsd"
    xmlns="http://www.example.com/example">
  <emma:one-of id="r1">
    <emma:interpretation id="int1">
      <origin>Boston</origin>
      <destination>Denver</destination>
      <date>03112003</date>
    </emma:interpretation>

    <emma:interpretation id="int2">
      <origin>Austin</origin>
      <destination>Denver</destination>
      <date>03112003</date>
    </emma:interpretation>
  </emma:one-of>
</emma:emma>

The function of the emma:one-of element is to represent a disjunctive list of possible interpretations of a user input. A disjunction of possible interpretations of an input can be the result of different kinds of processing or ambiguity. One source is multiple results from a recognition technology such as speech or handwriting recognition. Multiple results can also occur from parsing or understanding natural language. Another possible source of ambiguity is from the application of multiple different kinds of recognition or understanding components to the same input signal. For example, an single ink input signal might be processed by both handwriting recognition and gesture recognition. Another is the use of more than one recording device for the same input (multiple microphones).

In order to make explicit these different kinds of multiple interpretations and allow for concise statement of the annotations associated with each, the emma:one-of element MAY appear within another emma:one-of element. If emma:one-of elements are nested then they MUST indicate the kind of disjunction using the attribute disjunction-type. The values of disjunction-type are {recognition, understanding, multi-device, and multi-process}. For the most common use case, where there are multiple recognition results and some of them have multiple interpretations, the top-level emma:one-of is disjunction-type="recognition" and the embedded emma:one-of has the attribute disjunction-type="understanding".

As an example, in an interactive flight reservation application, recognition yielded 'Boston' or 'Austin' and each had a semantic interpretation as either the assertion of city name or the specification of a flight query with the city as the destination, this would be represented as follows in EMMA:


<emma:emma version="1.0"
    xmlns:emma="http://www.w3.org/2003/04/emma"
    xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
    xsi:schemaLocation="http://www.w3.org/2003/04/emma
     http://www.w3.org/TR/2007/WD-emma-20070409/emma.xsd"
    xmlns="http://www.example.com/example">
  <emma:one-of disjunction-type="recognition"
      start="12457990" end="12457995">
     <emma:one-of disjunction-type="understanding"
         emma:tokens="boston">
       <emma:interpretation>
          <assert><city>boston</city></assert>
       </emma:interpretation>
       <emma:interpretation>
          <flight><dest><city>boston</city></dest></flight>
       </emma:interpretation>
     </emma:one-of>
     <emma:one-of disjunction-type="understanding"
         emma:tokens="austin">
       <emma:interpretation>
          <assert><city>austin</city></assert>
       </emma:interpretation>
       <emma:interpretation>
          <flight><dest><city>austin</city></dest></flight>
       </emma:interpretation>
     </emma:one-of>
  </emma:one-of>
</emma:emma>

EMMA MAY explicitly represent ambiguity resulting from different processes, devices, or sources using embedded emma:one-of and the disjunction-type attribute. Multiple different interpretations resulting from different factors MAY also be listed within a single unstructured emma:one-of though in this case it is more complex or impossible to uncover the sources of the ambiguity if required by later stages of processing. If there is no embedding in emma:one-of, then the disjunction-type attribute is not required. If the disjunction-type attribute is missing then by default the source of disjunction is unspecified.

The example case above could also be represented as:


<emma:emma version="1.0"
    xmlns:emma="http://www.w3.org/2003/04/emma"
    xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
    xsi:schemaLocation="http://www.w3.org/2003/04/emma
     http://www.w3.org/TR/2007/WD-emma-20070409/emma.xsd"
    xmlns="http://www.example.com/example">
  <emma:one-of  start="12457990" end="12457995">
     <emma:interpretation emma:tokens="boston">
        <assert><city>boston</city></assert>
     </emma:interpretation>
     <emma:interpretation >
        <flight><dest><city>boston</city></dest></flight>
     </emma:interpretation>
     <emma:interpretation emma:tokens="austin">
        <assert><city>austin</city></assert>
     </emma:interpretation>
     <emma:interpretation emma:tokens="austin">
        <flight><dest><city>austin</city></dest></flight>
     </emma:interpretation>
  </emma:one-of>
</emma:emma>

But in this case information about which interpretations resulted from speech recognition and which resulted from language understanding is lost.

A list of emma:interpretation elements within an emma:one-of MUST be sorted best-first by some measure of quality. The quality measure is emma:confidence if present, otherwise, the quality metric is platform-specific.

With embedded emma:one-of structures there is no requirement for the confidence scores within different emma:one-of to be on the same scale. For example, the scores assigned by handwriting recognition might not be comparable to those assigned by gesture recognition. Similarly, if multiple recognizers are used there is no guarantee that their confidence scores will be comparable. For this reason the ordering requirement on emma:interpretation within emma:one-of only applies locally to sister emma:interpretation elements within each emma:one-of. There is no requirement on the ordering of embedded emma:one-of elements within a higher emma:one-of element.

3.3.2 emma:group element

Annotation emma:group
Definition A container element indicating that a number of interpretations of distinct user inputs are grouped according to some criteria.
Children The emma:group element MUST immediately contain a collection of one or more emma:interpretation elements or container elements: emma:one-of, emma:group, emma:sequence . It MAY also contain an optional single emma:group-info element. It MAY also contain multiple optional emma:derived-from elements and an optional single emma:info element.
Attributes
  • Required: Attribute id of type xsd:ID
  • Optional: The annotation attributes: emma:tokens, emma:process, emma:lang, emma:signal, emma:signal-size, emma:media-type, emma:confidence, emma:source, emma:start, emma:end, emma:time-ref-uri, emma:time-ref-anchor-point, emma:offset-to-start, emma:duration, emma:medium, emma:mode, emma:function, emma:verbal, emma:cost, emma:grammar-ref, emma:endpoint-info-ref, emma:model-ref, emma:dialog-turn.
Applies to The emma:group element is legal only as a child of emma:emma, emma:one-of, emma:group, emma:sequence, or emma:derivation.

The emma:group element is used to indicate that the contained interpretations are from distinct user inputs that are related in some manner. emma:group MUST NOT be used for containing the multiple stages of processing of a single user input. Those MUST be contained in the emma:derivation element instead (Section 4.1.2). For groups of inputs in temporal order the more specialized container emma:sequence MUST be used (Section 3.3.3). The following example shows three interpretations derived from the speech input "Move this ambulance here" and the tactile input related to two consecutive points on a map.

<emma:emma version="1.0"
    xmlns:emma="http://www.w3.org/2003/04/emma"
    xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
    xsi:schemaLocation="http://www.w3.org/2003/04/emma
     http://www.w3.org/TR/2007/WD-emma-20070409/emma.xsd"
    xmlns="http://www.example.com/example">
  <emma:group id="grp"
      emma:start="1087995961542"
      emma:end="1087995964542">
    <emma:interpretation id="int1">
      <action>move</action>
      <object>ambulance</object>
      <destination>here</destination>
    </emma:interpretation>

    <emma:interpretation id="int2">
      <x>0.253</x>
      <y>0.124</y>
    </emma:interpretation>

    <emma:interpretation id="int3">
      <x>0.866</x>
      <y>0.724</y>
    </emma:interpretation>
  </emma:group>
</emma:emma>

The emma:one-of and emma:group containers MAY be nested arbitrarily.

3.3.2.1 Indirect grouping criteria: emma:group-info element

Annotation emma:group-info
Definition The emma:group-info element contains or references criteria used in establishing the grouping of interpretations in an emma:group element.
Children The emma:group-info element MUST either immediately contain inline instance data specifying grouping criteria or have the attribute ref referencing the criteria.
Attributes
  • Optional: ref of type xsd:anyURI referencing the grouping criteria; alternatively the criteria MAY be provided inline as the content of the emma:group-info element.
Applies to The emma:group-info element is legal only as a child of emma:group.

Sometimes it may be convenient to indirectly associate a given group with information, such as grouping criteria. The emma:group-info element might be used to make explicit the criteria by which members of a group are associated. In the following example, a group of two points is associated with a description of grouping criteria based upon a sliding temporal window of two seconds duration.

<emma:emma version="1.0"
    xmlns:emma="http://www.w3.org/2003/04/emma"
    xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
    xsi:schemaLocation="http://www.w3.org/2003/04/emma
     http://www.w3.org/TR/2007/WD-emma-20070409/emma.xsd"
    xmlns="http://www.example.com/example"
    xmlns:ex="http://www.example.com/ns/group">
  <emma:group id="grp">
    <emma:group-info>
      <ex:mode>temporal</ex:mode>
      <ex:duration>2s</ex:duration>
    </emma:group-info>

    <emma:interpretation id="int1">
      <x>0.253</x>
      <y>0.124</y>
    </emma:interpretation>

    <emma:interpretation id="int2">
      <x>0.866</x>
      <y>0.724</y>
    </emma:interpretation>
  </emma:group>
</emma:emma>

You might also use emma:group-info to refer to a named grouping criterion using external reference, for instance:

<emma:emma version="1.0"
    xmlns:emma="http://www.w3.org/2003/04/emma"
    xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
    xsi:schemaLocation="http://www.w3.org/2003/04/emma
     http://www.w3.org/TR/2007/WD-emma-20070409/emma.xsd"
    xmlns="http://www.example.com/example"
    xmlns:ex="http://www.example.com/ns/group">
  <emma:group id="grp">
    <emma:group-info ref="http://www.example.com/criterion42"/>
    <emma:interpretation id="int1">
      <x>0.253</x>
      <y>0.124</y>
    </emma:interpretation>

    <emma:interpretation id="int2">
      <x>0.866</x>
      <y>0.724</y>
    </emma:interpretation>
  </emma:group>
</emma:emma>

3.3.3 emma:sequence element

Annotation emma:sequence
Definition A container element indicating that a number of interpretations of distinct user inputs are in temporal sequence.
Children The emma:sequence element MUST immediately contain a collection of one or more emma:interpretation elements or container elements: emma:one-of, emma:group, emma:sequence . It MAY also contain multiple optional emma:derived-from elements and an optional single emma:info element.
Attributes
  • Required: Attribute id of type xsd:ID
  • Optional: The annotation attributes: emma:tokens, emma:process, emma:lang, emma:signal, emma:signal-size, emma:media-type, emma:confidence, emma:source, emma:start, emma:end, emma:time-ref-uri, emma:time-ref-anchor-point, emma:offset-to-start, emma:duration, emma:medium, emma:mode, emma:function, emma:verbal, emma:cost, emma:grammar-ref, emma:endpoint-info-ref, emma:model-ref, emma:dialog-turn.
Applies to The emma:sequence element is legal only as a child of emma:emma, emma:one-of, emma:group, emma:sequence, or emma:derivation.

The emma:sequence element is used to indicate that the contained interpretations are sequential in time, as in the following example, which indicates that two points made with a pen are in temporal order.

<emma:emma version="1.0"
    xmlns:emma="http://www.w3.org/2003/04/emma"
    xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
    xsi:schemaLocation="http://www.w3.org/2003/04/emma
     http://www.w3.org/TR/2007/WD-emma-20070409/emma.xsd"
    xmlns="http://www.example.com/example">
  <emma:sequence id="seq1">
    <emma:interpretation id="int1" emma:medium="tactile"
        emma:mode="ink">
      <x>0.253</x>
      <y>0.124</y>
    </emma:interpretation>

    <emma:interpretation id="int2" emma:medium="tactile"
        emma:mode="ink">
      <x>0.866</x>
      <y>0.724</y>
    </emma:interpretation>
  </emma:sequence>
</emma:emma>

The emma:sequence container MAY be combined with emma:one-of and emma:group in arbitrary nesting structures. The order of children in the content of the emma:sequence element corresponds to a sequence of interpretations. This ordering does not imply any particular definition of sequentiality. EMMA processors are expected therefore to use the emma:sequence element to hold interpretations which are either strictly sequential in nature (e.g. the end-time of an interpretation precedes the start-time of its follower), or which overlap in some manner (e.g. the start-time of a follower interpretation precedes the end-time of its precedent). It is possible to use timestamps to provide fine grained annotation for the sequence of interpretations that are sequential in time (see Section 4.2.10).

In the following more complex example, a sequence of two pen gestures in emma:sequence and a speech input in emma:interpretation is contained in an emma:group.

<emma:emma version="1.0"
    xmlns:emma="http://www.w3.org/2003/04/emma"
    xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
    xsi:schemaLocation="http://www.w3.org/2003/04/emma
     http://www.w3.org/TR/2007/WD-emma-20070409/emma.xsd"
    xmlns="http://www.example.com/example">
  <emma:group id="grp">
     <emma:interpretation id="int1" emma:medium="acoustic"
         emma:mode="voice">
       <action>move</action>
       <object>this-battleship</object>
       <destination>here</destination>
     </emma:interpretation>

     <emma:sequence id="seq1">
       <emma:interpretation id="int2" emma:medium="tactile"
           emma:mode="ink">
         <x>0.253</x>
         <y>0.124</y>
       </emma:interpretation>

     <emma:interpretation id="int3" emma:medium="tactile"
         emma:mode="ink">
       <x>0.866</x>
       <y>0.724</y>
     </emma:interpretation>
   </emma:sequence>
 </emma:group>
</emma:emma>

3.4 Lattice element

In addition to providing the ability to represent N-best lists of interpretations using emma:one-of, EMMA also provides the capability to represent lattices of words or other symbols using the emma:lattice element. Lattices provide a compact representation of large lists of possible recognition results or interpretations for speech, pen, or multimodal inputs.

In addition to providing a representation for lattice output from speech recognition, another important use case for lattices is for representation of the results of gesture and handwriting recognition from a pen modality component. Lattices can also be used to compactly represent multiple possible meaning representations. Another use case for the lattice representation is for associating confidence scores and other annotations with individual words within a speech recognition result string.

Lattices are compactly described by a list of transitions between nodes. For each transition the start and end nodes MUST be defined, along with the label for the transition. Initial and final nodes MUST also be indicated. The following figure provides a graphical representation of a speech recognition lattice which compactly represents eight different sequences of words.

speech lattice

which expands to:

a. flights to boston from portland today please
b. flights to austin from portland today please
c. flights to boston from oakland today please
d. flights to austin from oakland today please
e. flights to boston from portland tomorrow
f. flights to austin from portland tomorrow
g. flights to boston from oakland tomorrow
h. flights to austin from oakland tomorrow

3.4.1 Lattice markup: emma:lattice, emma:arc, emma:node elements

Annotation emma:lattice
Definition An element which encodes a lattice representation of user input.
Children The emma:lattice element MUST immediately contain one or more emma:arc elements and zero or more emma:node elements.
Attributes
  • Required:
    • initial of type xsd:nonNegativeInteger indicating the number of the initial node of the lattice.
    • final contains a space-separated list of xsd:nonNegativeInteger indicating the numbers of the final nodes in the lattice.
  • Optional: emma:time-ref-uri, emma:time-ref-anchor-point.
Applies to The emma:lattice element is legal only as a child of the emma:interpretation element.
Annotation emma:arc
Definition An element which encodes a transition between two nodes in a lattice. The label associated with the arc in the lattice is represented in the content of emma:arc.
Children The emma:arc element MUST immediately contain either character data or a single application namespace element or be empty, in the case of epsilon transitions. It MAY contain an emma:info element containing application or vendor specific annotations.
Attributes
  • Required:
    • from of type xsd:nonNegativeInteger indicating the number of the starting node for the arc.
    • to of type xsd:nonNegativeInteger indicating the number of the ending node for the arc.
  • Optional: emma:start, emma:end, emma:offset-to-start, emma:duration, emma:confidence, emma:cost, emma:lang, emma:medium, emma:mode, emma:source.
Applies to The emma:arc element is legal only as a child of the emma:lattice element.
Annotation emma:node
Definition An element which represents a node in the lattice. The emma:node elements are not required to describe a lattice but might be added to provide a location for annotations on nodes in a lattice. There MUST be at most one emma:node specification for each numbered node in the lattice.
Children An OPTIONAL emma:info element for application or vendor specific annotations on the node.
Attributes
  • Required:
    • node-number of type xsd:nonNegativeInteger indicating the node number in the lattice.
  • Optional: emma:start, emma:end, emma:offset-to-start, emma:confidence, emma:cost.
Applies to The emma:node element is legal only as a child of the emma:lattice element.

In EMMA, a lattice is represented using an element emma:lattice, which has attributes initial and final for indicating the initial and final nodes of the lattice. For the lattice below, this will be: <emma:lattice initial="1" final="8"/>. The nodes are numbered with integers. If there is more than one distinct final node in the lattice the nodes MUST be represented as a space separated list in the value of the final attribute e.g. <emma:lattice initial="1" final="9 10 23"/>. There MUST only be one initial node in an EMMA lattice. Each transition in the lattice is represented as an element emma:arc with attributes from and to which indicate the nodes where the transition starts and ends. The arc's label is represented as the content of the emma:arc element and MUST be any well-formed character or XML content. In the example here the contents are words. Empty (epsilon) transitions in a lattice MUST be represented in the emma:lattice representation as emma:arc empty elements, e.g. <emma:arc from="1" to="8"/>.

The example speech lattice above would be represented in EMMA markup as follows:

<emma:emma version="1.0"
    xmlns:emma="http://www.w3.org/2003/04/emma"
    xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
    xsi:schemaLocation="http://www.w3.org/2003/04/emma
     http://www.w3.org/TR/2007/WD-emma-20070409/emma.xsd"
    xmlns="http://www.example.com/example">
  <emma:interpretation id="interp1">
    <emma:lattice initial="1" final="8">
      <emma:arc from="1" to="2">flights</emma:arc>

      <emma:arc from="2" to="3">to</emma:arc>
      <emma:arc from="3" to="4">boston</emma:arc>
      <emma:arc from="3" to="4">austin</emma:arc>
      <emma:arc from="4" to="5">from</emma:arc>

      <emma:arc from="5" to="6">portland</emma:arc>
      <emma:arc from="5" to="6">oakland</emma:arc>
      <emma:arc from="6" to="7">today</emma:arc>
      <emma:arc from="7" to="8">please</emma:arc>

      <emma:arc from="6" to="8">tomorrow</emma:arc>
    </emma:lattice>
  </emma:interpretation>
</emma:emma>

Alternatively, if we wish to represent the same information as an N-best list using emma:one-of, we would have the more verbose representation:

<emma:emma version="1.0"
    xmlns:emma="http://www.w3.org/2003/04/emma"
    xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
    xsi:schemaLocation="http://www.w3.org/2003/04/emma
     http://www.w3.org/TR/2007/WD-emma-20070409/emma.xsd"
    xmlns="http://www.example.com/example">
  <emma:one-of id="nbest1">
    <emma:interpretation id="interp1">
      <text>flights to boston from portland today please</text>
    </emma:interpretation>

    <emma:interpretationid="interp2">
      <text>flights to boston from portland tomorrow</text>
    </emma:interpretation>

    <emma:interpretation id="interp3">
      <text>flights to austin from portland today please</text>
    </emma:interpretation>

    <emma:interpretation id="interp4">
      <text>flights to austin from portland tomorrow</text>
    </emma:interpretation>

    <emma:interpretation id="interp5">
      <text>flights to boston from oakland today please</text>
    </emma:interpretation>

    <emma:interpretation id="interp6">
      <text>flights to boston from oakland tomorrow</text>
    </emma:interpretation>

    <emma:interpretation id="interp7">
      <text>flights to austin from oakland today please</text>
    </emma:interpretation>

    <emma:interpretation id="interp8">
      <text>flights to austin from oakland tomorrow</text>
    </emma:interpretation>
  </emma:one-of>
</emma:emma>

The lattice representation avoids the need to enumerate all of the possible word sequences. Also, as detailed below, the emma:lattice representation enables placement of annotations on individual words in the input.

For use cases involving the representation of gesture/ink lattices and use cases involving lattices of semantic interpretations, EMMA allows for application namespace elements to appear within emma:arc.

For example a sequence of two gestures, each of which is recognized as either a line or a circle, might be represented as follows:

<emma:emma version="1.0"
    xmlns:emma="http://www.w3.org/2003/04/emma"
    xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
    xsi:schemaLocation="http://www.w3.org/2003/04/emma
     http://www.w3.org/TR/2007/WD-emma-20070409/emma.xsd"
    xmlns="http://www.example.com/example">
  <emma:interpretation id="interp1">
    <emma:lattice initial="1" final="3">
      <emma:arc from="1" to="2">
        <circle radius="100"/>
      </emma:arc>
      <emma:arc from="2" to="3">
        <line length="628"/>
      </emma:arc>
      <emma:arc from="1" to="2">
        <circle radius="200"/>
      </emma:arc>
      <emma:arc from="2" to="3">
        <line length="1256"/>
      </emma:arc>
    </emma:lattice>
  </emma:interpretation>
</emma:emma>

As an example of a lattice of semantic interpretations, in a travel application where the source is either "Boston" or "Austin"and the destination is either "Newark" or "New York", the possibilities might be represented in a lattice as follows:

<emma:emma version="1.0"
    xmlns:emma="http://www.w3.org/2003/04/emma"
    xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
    xsi:schemaLocation="http://www.w3.org/2003/04/emma
     http://www.w3.org/TR/2007/WD-emma-20070409/emma.xsd"
    xmlns="http://www.example.com/example">
  <emma:interpretation id="interp1">
    <emma:lattice initial="1" final="3">
      <emma:arc from="1" to="2">
        <source city="boston"/>
      </emma:arc>
      <emma:arc from="2" to="3">
        <destination city="newark"/>
      </emma:arc>
      <emma:arc from="1" to="2">
        <source city="austin"/>
      </emma:arc>
      <emma:arc from="2" to="3">
        <destination city="new york"/>
      </emma:arc>
    </emma:lattice>
  </emma:interpretation>
</emma:emma>

The emma:arc element MAY contain either an application namespace element or character data. It MUST NOT contain combinations of application namespace elements and character data. However, an emma:info element MAY appear within an emma:arc element alongside character data, in order to allow for the association of vendor or application specific annotations on a single word or symbol in a lattice.

So, in summary, there are four groupings of content that can appear within emma:arc:

3.4.2 Annotations on lattices

The encoding of lattice arcs as XML elements (emma:arc) enables arcs to be annotated with metadata such as timestamps, costs, or confidence scores:

<emma:emma version="1.0"
    xmlns:emma="http://www.w3.org/2003/04/emma"
    xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
    xsi:schemaLocation="http://www.w3.org/2003/04/emma
     http://www.w3.org/TR/2007/WD-emma-20070409/emma.xsd"
    xmlns="http://www.example.com/example">
  <emma:interpretation id="interp1">
    <emma:lattice initial="1" final="8">
      <emma:arc
       from="1"
       to="2"
       emma:start="1087995961542"
       emma:end="1087995962042"
       emma:cost="30">
         flights
      </emma:arc>

      <emma:arc
       from="2"
       to="3"
       emma:start="1087995962042"
       emma:end="1087995962542"
       emma:cost="20">
         to
      </emma:arc>

      <emma:arc
       from="3"
       to="4"
       emma:start="1087995962542"
       emma:end="1087995963042"
       emma:cost="50">
         boston
      </emma:arc>

      <emma:arc
       from="3"
       to="4"
       emma:start="1087995963042"
       emma:end="1087995963742"
       emma:cost="60">
         austin
      </emma:arc>
      ...
    </emma:lattice>
  </emma:interpretation>
</emma:emma>

The following EMMA attributes MAY be placed on emma:arc elements: absolute timestamps (emma:start, emma:end), relative timestamps ( emma:offset-to-start, emma:duration), emma:confidence, emma:cost, the human language of the input (emma:lang), emma:medium, emma:mode, and emma:source. The use case for emma:medium, emma:mode, and emma:source is for lattices which contains content from different input modes. The emma:arc element MAY also contain an emma:info element for specification of vendor and application specific annotations on the arc.

Costs are typically application and device dependent. There are a variety of ways that individual arc costs might be combined to produce costs for specific paths through the lattice. This specification does not standardize the way for these costs to be combined; it is up to the applications and devices to determine how such derived costs would be computed and used.

For some lattice formats, it is also desirable to annotate the nodes in the lattice themselves with information such as costs. For example in speech recognition, costs might be placed on nodes as a result of word penalities or redistribution of costs. For this purpose EMMA also provides an emma:node element which can host annotations such as emma:cost. The emma:node element MUST have an attribute node-number which indicates the number of the node. There MUST be at most one emma:node specification for a given numbered node in the lattice. In our example, if there was a cost of 100 on the final state this could be represented as follows:

<emma:emma version="1.0"
    xmlns:emma="http://www.w3.org/2003/04/emma"
    xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
    xsi:schemaLocation="http://www.w3.org/2003/04/emma
     http://www.w3.org/TR/2007/WD-emma-20070409/emma.xsd"
    xmlns="http://www.example.com/example">
  <emma:interpretation id="interp1">
    <emma:lattice initial="1" final="8">
      <emma:arc
       from="1"
       to="2"
       emma:start="1087995961542"
       emma:end="1087995962042"
       emma:cost="30">
         flights
      </emma:arc>
      <emma:arc
       from="2"
       to="3"
       emma:start="1087995962042"
       emma:end="1087995962542"
       emma:cost="20">
         to
      </emma:arc>

      <emma:arc
       from="3"
       to="4"
       emma:start="1087995962542"
       emma:end="1087995963042"
       emma:cost="50">
         boston
      </emma:arc>
      <emma:arc
       from="3"
       to="4"
       emma:start="1087995963042"
       emma:end="1087995963742"
       emma:cost="60">
         austin
      </emma:arc>
        ...
      <emma:node node-number="8" emma:cost="100"/>
    </emma:lattice>
  </emma:interpretation>
</emma:emma>

3.4.3 Relative timestamps on lattices

The relative timestamp mechanism in EMMA is intended to provide temporal information about arcs in a lattice in relative terms using offsets in milliseconds. In order to do this the absolute time MAY be specified on emma:interpretation; both emma:time-ref-uri and emma:time-ref-anchor-point apply to emma:lattice and MAY be used there to set the anchor point for offsets to the start of the absolute time specified on emma:interpretation. The offset in milliseconds to the beginning of each arc MAY then be indicated on each emma:arc in the emma:offset-to-start attribute.

<emma:emma version="1.0"
    xmlns:emma="http://www.w3.org/2003/04/emma"
    xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
    xsi:schemaLocation="http://www.w3.org/2003/04/emma
     http://www.w3.org/TR/2007/WD-emma-20070409/emma.xsd"
    xmlns="http://www.example.com/example">

  <emma:interpretation id="interp1"
          emma:start="1087995961542" emma:end="1087995963042">
    <emma:lattice emma:time-ref-uri="#interp1"
        emma:time-ref-anchor-point="start"
        initial="1" final="4">
      <emma:arc
       from="1"
       to="2"
       emma:offset-to-start="0">
         flights
      </emma:arc>
      <emma:arc
       from="2"
       to="3"
       emma:offset-to-start="500">
         to
      </emma:arc>

      <emma:arc
       from="3"
       to="4"
       emma:offset-to-start="1000">
         boston
      </emma:arc>
    </emma:lattice>
  </emma:interpretation>
</emma:emma>

Note that the offset for the first emma:arc MUST always be zero since the EMMA attribute emma:offset-to-start indicates the number of milliseconds from the anchor point to the start of the piece of input associated with the emma:arc, in this case the word "flights".

3.5 Literal semantics: emma:literal element

Annotation emma:literal
Definition An element that contains string literal output.
Children String literal
Attributes None.
Applies to The emma:literal is a child of emma:interpretation.

Certain EMMA processing components produce semantic results in the form of string literals without any surrounding application namespace markup. These MUST be placed with the EMMA element emma:literal within emma:interpretation. For example, if a semantic interpreter simply returned "boston" this could be represented in EMMA as:

<emma:emma version="1.0"
    xmlns:emma="http://www.w3.org/2003/04/emma"
    xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
    xsi:schemaLocation="http://www.w3.org/2003/04/emma
     http://www.w3.org/TR/2007/WD-emma-20070409/emma.xsd"
    xmlns="http://www.example.com/example">
  <emma:interpretation>
    <emma:literal>boston</emma:literal>
  </emma:interpretation>
</emma:emma>

4. EMMA annotations

This section defines annotations in the EMMA namespace including both attributes and elements. The values are specified in terms of the data types defined by XML Schema Part 2: Datatypes Second Edition [XML Schema Datatypes].

4.1 EMMA annotation elements

4.1.1 Data model: emma:model element

Annotation emma:model
Definition The emma:model either references or provides inline the data model for the instance data.
Children If a ref attribute is not specified then this element contains the data model inline.
Attributes
  • Required:
    • id of type xsd:ID.
  • Optional:
    • ref of type xsd:anyURI that references the data model. Note that either an ref attribute or in-line data model (but not both) MUST be specified.
Applies to The emma:model element MAY appear only as a child of emma:emma.

The data model that may be used to express constraints on the structure and content of instance data is specified as one of the annotations of the instance. Specifying the data model is OPTIONAL, in which case the data model can be said to be implicit. Typically the data model is pre-established by the application.

The data model is specified with the emma:model annotation defined as an element in the EMMA namespace. If the data model for the contents of a emma:interpretation, container elements, or application namespace element is to be specified in EMMA, the attribute emma:model-ref MUST be specified on the emma:interpretation, container element, or application namespace element. Note that since multiple emma:model elements might be specified under the emma:emma it is possible to refer to multiple data models within a single EMMA document. For example, different alternative interpretations under an emma:one-of might have different data models. In this case, an emma:model-ref attribute would appear on each emma:interpretation element in the N-best list with its value being the id of the emma:model element for that particular interpretation.

The data model is closely related to the interpretation data, and is typically specified as the annotation related to the emma:interpretation or emma:one-of elements.

<emma:emma version="1.0"
    xmlns:emma="http://www.w3.org/2003/04/emma"
    xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
    xsi:schemaLocation="http://www.w3.org/2003/04/emma
     http://www.w3.org/TR/2007/WD-emma-20070409/emma.xsd"
    xmlns="http://www.example.com/example">
  <emma:model id="model1" ref="http://example.com/models/city.xml"/>
  <emma:interpretation id="int1" emma:model-ref="model1">
    <city> London </city>
    <country> UK </country>
  </emma:interpretation>
</emma:emma>

The emma:model annotation MAY reference any element or attribute in the application instance data, as well as any EMMA container element (emma:one-of, emma:group, or emma:sequence).

The data model annotation MAY be used to either reference an external data model with the ref attribute or provide a data model as in-line content. Either a ref attribute or in-line data model (but not both) MUST be specified.

4.1.2 Interpretation derivation: emma:derived-from element and emma:derivation element

Annotation emma:derived-from
Definition An empty element which provides a reference to the interpretation which the element it appears on was derived from.
Children None
Attributes
  • Required:
    • resource of type xsd:anyURI that references the interpretation from which the current interpretation is derived.
  • Optional:
    • composite of type xsd:boolean that is "true" if the derivation step combines multiple inputs and "false" if not. If composite is not specified the value is "false" by default.
Applies to The emma:derived-from element is legal only as a child of emma:interpretation, emma:one-of, emma:group, or emma:sequence.
Annotation emma:derivation
Definition An element which contains interpretation and container elements representing earlier stages in the processing of the input.
Children One or more emma:interpretation, emma:one-of, emma:sequence, or emma:group elements.
Attributes None
Applies to The emma:derivation MAY appear only as a child of the emma:emma element.

Instances of interpretations are in general derived from other instances of interpretation in a process that goes from raw data to increasingly refined representations of the input. The derivation annotation is used to link any two interpretations that are related by representing the source and the outcome of an interpretation process. For instance, a speech recognition process can return the following result in the form of raw text:

<emma:emma version="1.0"
    xmlns:emma="http://www.w3.org/2003/04/emma"
    xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
    xsi:schemaLocation="http://www.w3.org/2003/04/emma
     http://www.w3.org/TR/2007/WD-emma-20070409/emma.xsd"
    xmlns="http://www.example.com/example">
  <emma:interpretation id="raw">
    <answer>From Boston to Denver tomorrow</answer>
  </emma:interpretation>

</emma:emma>

A first interpretation process will produce:

<emma:emma version="1.0"
    xmlns:emma="http://www.w3.org/2003/04/emma"
    xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
    xsi:schemaLocation="http://www.w3.org/2003/04/emma
     http://www.w3.org/TR/2007/WD-emma-20070409/emma.xsd"
    xmlns="http://www.example.com/example">
  <emma:interpretation id="better">
    <origin>Boston</origin>
    <destination>Denver</destination>
    <date>tomorrow</date>
  </emma:interpretation>
</emma:emma>

A second interpretation process, aware of the current date, will be able to produce a more refined instance, such as:

<emma:emma version="1.0"
    xmlns:emma="http://www.w3.org/2003/04/emma"
    xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
    xsi:schemaLocation="http://www.w3.org/2003/04/emma
     http://www.w3.org/TR/2007/WD-emma-20070409/emma.xsd"
    xmlns="http://www.example.com/example">
  <emma:interpretation id="best">
    <origin>Boston</origin>
    <destination>Denver</destination>
    <date>20030315</date>
  </emma:interpretation>
</emma:emma>

The interaction manager might need to have access to the three levels of interpretation. The emma:derived-from annotation element can be used to establish a chain of derivation relationships as in the following example:

<emma:emma version="1.0"
    xmlns:emma="http://www.w3.org/2003/04/emma"
    xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
    xsi:schemaLocation="http://www.w3.org/2003/04/emma
     http://www.w3.org/TR/2007/WD-emma-20070409/emma.xsd"
    xmlns="http://www.example.com/example">
  <emma:derivation>
    <emma:interpretation id="raw">
      <answer>From Boston to Denver tomorrow</answer>
    </emma:interpretation>

    <emma:interpretation id="better">
      <emma:derived-from resource="#raw" composite="false"/>
      <origin>Boston</origin>
      <destination>Denver</destination>
      <date>tomorrow</date>
    </emma:interpretation>
  </emma:derivation>

  <emma:interpretation id="best">
    <emma:derived-from resource="#better" composite="false"/>
    <origin>Boston</origin>
    <destination>Denver</destination>
    <date>20030315</date>
  </emma:interpretation>
</emma:emma>

The emma:derivation element MAY be used as a container for representations of the earlier stages in the interpretation of the input. The latest stage of processing MUST be a direct child of emma:emma.

The resource attribute on emma:derived-from is a URI which can reference IDs in the current or other EMMA documents.

In addition to representing sequential derivations, the EMMA emma:derived-from element can also be used to capture composite derivations. Composite derivations involve combination of inputs from different modes.

In order to indicate whether an emma:derived-from element describes a sequential derivation step or a composite derivation step, the emma:derived-from element has an attribute composite which has a boolean value. A composite emma:derived-from MUST be marked as composite="true" while a sequential emma:derived-from element is marked as composite="false". If this attribute is not specified the value is false by default.

In the following composite derivation example the user said "destination" using the voice mode and circled Boston on a map using the ink mode:

<emma:emma version="1.0"
    xmlns:emma="http://www.w3.org/2003/04/emma"
    xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
    xsi:schemaLocation="http://www.w3.org/2003/04/emma
     http://www.w3.org/TR/2007/WD-emma-20070409/emma.xsd"
    xmlns="http://www.example.com/example">
  <emma:derivation>
    <emma:interpretation id="voice1"
        emma:start="1087995961500"
        emma:end="1087995962542"
        emma:process="http://example.com/myasr.xml"
        emma:source="http://example.com/microphone/NC-61"
        emma:signal="http://example.com/signals/sg23.wav"
        emma:confidence="0.6"
        emma:medium="acoustic"
        emma:mode="voice"
        emma:function="dialog"
        emma:verbal="true"
        emma:lang="en-US"
        emma:tokens="destination">
      <rawinput>destination</rawinput>
    </emma:interpretation>

    <emma:interpretation id="ink1"
        emma:start="1087995961600"
        emma:end="1087995964000"
        emma:process="http://example.com/mygesturereco.xml"
        emma:source="http://example.com/pen/wacom123"
        emma:signal="http://example.com/signals/ink5.inkml"
        emma:confidence="0.5"
        emma:medium="tactile"
        emma:mode="ink"
        emma:function="dialog"
        emma:verbal="false">
      <rawinput>Boston</rawinput>
    </emma:interpretation>
  </emma:derivation>

  <emma:interpretation id="multimodal1"
      
      
      emma:confidence="0.3"
      emma:start="1087995961500"
      emma:end="1087995964000"
      emma:medium="acoustic tactile"
      emma:mode="voice ink"
      emma:function="dialog"
      emma:verbal="true"
      emma:lang="en-US"
      emma:tokens="destination">
    <emma:derived-from resource="#voice1" composite="true"
    <emma:derived-from resource="#ink1" composite="true"
    <destination>Boston</destination>
  </emma:interpretation>
</emma:emma>

In this example, annotations on the multimodal interpretation indicate the process used for the integration and there are two emma:derived-from elements, one pointing to the speech and one pointing to the pen gesture.

The only constraints the EMMA specification places on the annotations that appear on a composite input are that the emma:medium attribute MUST contain the union of the emma:medium attributes on the combining inputs, represented as a space delimited set of nmtokens as defined in Section 4.3, and that the emma:mode attribute MUST contain the union of the emma:mode attributes on the combining inputs, represented as a space delimited set of tokens. In the example above this meanings that the emma:medium value is "acoustic tactile" and the emma:mode attribute is "voice ink". How all other annotations are handled is author defined. In the following paragraph, informative examples on how specific annotations might be handled are given.

With reference to the illustrative example above, this paragraph provides informative guidance regarding the determination of annotations (beyond emma:medium and emma:mode on a composite multimodal interpretation). Generally the timestamp on a combined input should contain the intervals indicated by the combining inputs. For the absolute timestamps emma:start and emma:end this can be achieved by taking the earlier of the emma:start values (emma:start="1087995961500" in our example) and the later of the emma:end values (emma:end="1087995964000" in the example). The determination of relative timestamps for composite is more complex, informative guidance is given in Section 4.2.10.4. Generally speaking the emma:confidence value will be some numerical combination of the confidence scores assigned to the combining inputs. In our example, it is the result of multiplying the voice and ink confidence scores (0.3). In other cases there may not be a confidence score for one of the combining inputs and the author may choose to copy the confidence score from the input which does have one. Generally, for emma:verbal, if either of the inputs has the value true then the multimodal interpretation will also be emma:verbal="true" as in the example. In other words the annotation for the composite input is the result of an inclusive OR of the boolean values of the annotations on the inputs. If an annotation is only specified on one of the combining inputs then it may in some cases be assumed to apply to the multimodal interpretation of the composite input. In the example, emma:lang="en-US" is only specified for the speech input, and this annotation appears on the composite result also. Similarly in our example, only the voice has emma:tokens and the author has chosen to annotate the combined input with the same emma:tokens value. In this example, the emma:function is the same on both combining input and the author has chosen to use the same annotation on the composite interpretation.

In annotating derivations of the processing of the input, EMMA provides the flexibility of both course-grained or fine-grained annotation of relations among interpretations. For example, when relating two N-best lists, within emma:one-of elements either there can be a single emma:derived-from element under emma:one-of referring to the ID of the emma:one-of for the earlier processing stage:


<emma:emma version="1.0"
    xmlns:emma="http://www.w3.org/2003/04/emma"
    xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
    xsi:schemaLocation="http://www.w3.org/2003/04/emma
     http://www.w3.org/TR/2007/WD-emma-20070409/emma.xsd"
    xmlns="http://www.example.com/example">
  <emma:derivation>
    <emma:one-of id="nbest1">
      <emma:interpretation id="int1">
       <res>from boston to denver on march eleven two thousand three</res>
      </emma:interpretation>

      <emma:interpretation id="int2">
       <res>from austin to denver on march eleven two thousand three</res>
      </emma:interpretation>
  </emma:one-of>
</emma:derivation>

<emma:one-of id="nbest2">
  <emma:derived-from resource="#nbest1" composite="false"/>
  <emma:interpretation id="int1b">
    <origin>Boston</origin>
    <destination>Denver</destination>
    <date>03112003</date>
  </emma:interpretation>

  <emma:interpretation id="int2b">
    <origin>Austin</origin>
    <destination>Denver</destination>
    <date>03112003</date>
  </emma:interpretation>
</emma:one-of>
  
</emma:emma>

Or there can be a separate emma:derived-from element on each emma:interpretation element referring to the specific emma:interpretation element it was derived from.

<emma:emma version="1.0"
    xmlns:emma="http://www.w3.org/2003/04/emma"
    xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
    xsi:schemaLocation="http://www.w3.org/2003/04/emma
     http://www.w3.org/TR/2007/WD-emma-20070409/emma.xsd"
    xmlns="http://www.example.com/example">
  <emma:one-of id="nbest2">
    <emma:interpretation id="int1b">
     <emma:derived-from resource="#int1" composite="false"/>
      <origin>Boston</origin>
      <destination>Denver</destination>
      <date>03112003</date>
    </emma:interpretation>

    <emma:interpretation id="int2b">
     <emma:derived-from resource="#int2" composite="false"/>
      <origin>Austin</origin>
      <destination>Denver</destination>
      <date>03112003</date>
    </emma:interpretation>
  </emma:one-of>
  <emma:derivation>
    <emma:one-of id="nbest1">
      <emma:interpretation id="int1">
       <res>from boston to denver on march eleven two thousand three</res>
      </emma:interpretation>

      <emma:interpretation id="int2">
       <res>from austin to denver on march eleven two thousand three</res>
      </emma:interpretation>
    </emma:one-of>
  </emma:derivation>
</emma:emma>

Section 4.3 provides further examples of the use of emma:derived-from to represent sequential derivations and addresses the issue of the scope of EMMA annotations across derivations of user input.

4.1.3 Reference to grammar used: emma:grammar element

Annotation emma:grammar
Definition An element used to provide a reference to the grammar used in processing the input.
Children None
Attributes
  • Required:
    • ref of type xsd:anyURI that references a grammar used in processing the input.
    • id of type xsd:ID.
Applies to The emma:grammar is legal only as a child of the emma:emma element.

The grammar that was used to derive the EMMA result MAY be specified with the emma:grammar annotation defined as an element in the EMMA namespace.

Example:

<emma:emma version="1.0"
    xmlns:emma="http://www.w3.org/2003/04/emma"
    xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
    xsi:schemaLocation="http://www.w3.org/2003/04/emma
     http://www.w3.org/TR/2007/WD-emma-20070409/emma.xsd"
    xmlns="http://www.example.com/example">
  <emma:grammar id="gram1" ref="someURI"/>
  <emma:grammar id="gram2" ref="anotherURI"/>
  <emma:one-of id="r1">
    <emma:interpretation id="int1" emma:grammar-ref="gram1">
      <origin>Boston</origin>
    </emma:interpretation>

    <emma:interpretation id="int2" emma:grammar-ref="gram1">
        <origin>Austin</origin>
    </emma:interpretation>

    <emma:interpretation id="int3" emma:grammar-ref="gram2">
        <command>help</command>
    </emma:interpretation>
  </emma:one-of>
</emma:emma>

The emma:grammar annotation is a child of emma:emma.

4.1.4 Extensibility to application/vendor specific annotations: emma:info element

Annotation emma:info
Definition The emma:info element acts as a container for vendor and/or application specific metadata regarding a user's input.
Children One of more elements in the application namespace providing metadata about the input.
Attributes
  • Optional:
    • id of type xsd:ID.
Applies to The emma:info element is legal only as a child of the EMMA elements emma:emma, emma:interpretation, emma:group, emma:one-of, emma:sequence, emma:arc, or emma:node.

In Section 4.2, a series of attributes are defined for representation of metadata about user inputs in a standardized form. EMMA also provides an extensibility mechanism for annotation of user inputs with vendor or application specific metadata not covered by the standard set of EMMA annotations. The element emma:info MUST be used as a container for these annotations, UNLESS they are explicitly covered by emma:endpoint-info. For example, if an input to a dialog system needed to be annotated with the number that the call originated from, their state, some indication of the type of customer, and the name of the service, these pieces of information could be represented within emma:info as in the following example:

<emma:emma version="1.0"
    xmlns:emma="http://www.w3.org/2003/04/emma"
    xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
    xsi:schemaLocation="http://www.w3.org/2003/04/emma
     http://www.w3.org/TR/2007/WD-emma-20070409/emma.xsd"
    xmlns="http://www.example.com/example">
  <emma:info>
    <caller_id>
      <phone_number>2121234567</phone_number>
      <state>NY</state>
    </caller_id>

    <customer_type>residential</customer_type>
    <service_name>acme_travel_service</service_name>
  </emma:info>

  <emma:one-of id="r1" emma:start="1087995961542"
      emma:end="1087995963542">
    <emma:interpretation id="int1" emma:confidence="0.75">
      <origin>Boston</origin>
      <destination>Denver</destination>
      <date>03112003</date>
    </emma:interpretation>

    <emma:interpretation id="int2" emma:confidence="0.68">
      <origin>Austin</origin>
      <destination>Denver</destination>
      <date>03112003</date>
    </emma:interpretation>
  </emma:one-of>
</emma:emma>

It is important to have an EMMA container element for application/vendor specific annotations since EMMA elements provide a structure for representation of multiple possible interpretations of the input. As a result it is cumbersome to state application/vendor specific metadata as part of the application data within each emma:interpretation. An element is used rather than an attribute so that internal structure can be given to the annotations within emma:info.

In addition to emma:emma, emma:info MAY also appear as a child of other structural elements such as emma:interpretation, emma:info and so on. When emma:info appears as a child of one of these elements the application/vendor specific annotations contained within emma:info are assumed to apply to all of the emma:interpretation elements within the containing element. The semantics of conflicting annotations in emma:info, for example when different values are found within emma:emma and emma:interpretation, are left to the developer of the vendor/application specific annotations.

4.1.5 Endpoint reference: emma:endpoint-info element and emma:endpoint element

Annotation emma:endpoint-info
Definition The emma:endpoint-info element acts as a container for all application specific annotation regarding the communication environment.
Children One or more emma:endpoint elements.
Attributes
  • Required:
    • id of type xsd:ID.
Applies to The emma:endpoint-info elements is legal only as a child of emma:emma.
Annotation emma:endpoint
Definition The element acts as a container for application specific endpoint information.
Children Elements in the application namespace providing metadata about the input.
Attributes
  • Required:
    • id of type xsd:ID
  • Optional: emma:endpoint-role, emma:endpoint-address, emma:message-id, emma:port-num, emma:port-type, emma:endpoint-pair-ref, emma:service-name, emma:media-type, emma:medium, emma:mode.
Applies to emma:endpoint-info

In order to conduct multimodal interaction, there is a need in EMMA to specify the properties of the endpoint that receives the input which leads to the EMMA annotation. This allows subsequent components to utilize the endpoint properties as well as the annotated inputs to conduct meaningful multimodal interaction. EMMA element emma:endpoint can be used for this purpose. It can specify the endpoint properties based on a set of common endpoint property attributes in EMMA, such as emma:endpoint-address, emma:port-num, emma:port-type, etc. (Section 4.2.14). Moreover, it provides an extensible annotation structure that allows the inclusion of application and vendor specific endpoint properties.

Note that the usage of the term "endpoint" in this context is different from the way that the term is used in speech processing, where it refers to the end of a speech input. As used here, "endpoint" refers to a network location which is the source or receipient of an EMMA document.

In multimodal interaction, multiple devices can be used and each device can open multiple communication endpoints at the same time. These endpoints are used to transmit and receive data, such as raw input, EMMA documents, etc. The EMMA element emma:endpoint provides a generic representation of endpoint information which is relevant to multimodal interaction. It allows the annotation to be interoperable, and it eliminates the need for EMMA processors to create their own specialized annotations for existing protocols, potential protocols or yet undefined private protocols that they may use.

Moreover, emma:endpoint-info provides a container to hold all annotations regarding the endpoint information, including emma:endpoint and other application and vendor specific annotations that are related to the communication, allowing the same communication environment to be referenced and used in multiple interpretations.

Note that EMMA provides two locations (i.e. emma:info and emma:endpoint-info) for specifying vendor/application specific annotations. If the annotation is specifically related to the description of the endpoint, then the vendor/application specific annotation SHOULD be placed within emma:endpoint-info, otherwise it SHOULD be placed within emma:info.

The following example illustrates the annotation of endpoint reference properties in EMMA.

<emma:emma version="1.0"
    xmlns:emma="http://www.w3.org/2003/04/emma"
    xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
    xsi:schemaLocation="http://www.w3.org/2003/04/emma
     http://www.w3.org/TR/2007/WD-emma-20070409/emma.xsd"
    xmlns="http://www.example.com/example"
    xmlns:ex="http://www.example.com/emma/port">
  <emma:endpoint-info id="audio-channel-1">
    <emma:endpoint id="endpoint1"
        emma:endpoint-role="sink"
        emma:endpoint-address="135.61.71.103"
        emma:port-num="50204"
        emma:port-type="rtp"
        emma:endpoint-pair-ref="endpoint2"
        emma:media-type="audio/dsr-202212; rate:8000; maxptime:40"
        emma:service-name="travel"
        emma:mode="voice">
      <ex:app-protocol>SIP</ex:app-protocol>
    </emma:endpoint>

    <emma:endpoint id="endpoint2"
        emma:endpoint-role="source"
        emma:endpoint-address="136.62.72.104"
        emma:port-num="50204"
        emma:port-type="rtp"
        emma:endpoint-pair-ref="endpoint1"
        emma:media-type="audio/dsr-202212; rate:8000; maxptime:40"
        emma:service-name="travel"
        emma:mode="voice">
      <ex:app-protocol>SIP</ex:app-protocol>
    </emma:endpoint>
  </emma:endpoint-info>

  <emma:interpretation id="int1"
      emma:start="1087995961542" emma:end="1087995963542"
      emma:endpoint-info-ref="audio-channel-1">
    <destination>Chicago</destination>
  </emma:interpretation>
</emma:emma>

The ex:app-protocol is provided by the application or the vendor specification. It specifies that the application layer protocol used to establish the speech transmission from the "source" port to the "sink" port is Session Initiation Protocol (SIP). This is specific to SIP based VoIP communication, in which the actual media transmission and the call signaling that controls the communication sessions, are separated and typically based on different protocols. In the above example, the Real-time Transmission Protocol (RTP) is used in the media transmission between the source port and the sink port.

4.2 EMMA annotation attributes

4.2.1 Tokens of input: emma:tokens attribute

Annotation emma:tokens
Definition An attribute of type xsd:string holding a sequence of input tokens.
Applies to emma:interpretation, emma:group, emma:one-of, emma:sequence, and application instance data.

The emma:tokens annotation holds a list of input tokens. In the following description, the term tokens is used in the computational and syntactic sense of units of input, and not in the sense of XML tokens. The value held in emma:tokens is the list of the tokens of input as produced by the processor which generated the EMMA document; there is no language associated with this value.

In the case where a grammar is used to constrain input, the value will correspond to tokens as defined by the grammar. So for an EMMA document produced by input to a SRGS grammar [SRGS], the value of emma:tokens will be the list of words and/or phrases that are defined as tokens in SRGS (see Section 2.1 of [SRGS]). Items in the emma:tokens list are delimited by white space and/or quotation marks for phrases containing white space. For example:

emma:tokens="arriving at 'Liverpool Street'"

where the three tokens of input are arriving, at and Liverpool Street.

The emma:tokens annotation MAY be applied not just to the lexical words and phrases of language but to any level of input processing. Other examples of tokenization include phonemes, ink strokes, gestures and any other discrete units of input at any level.

Examples:

<emma:emma version="1.0"
    xmlns:emma="http://www.w3.org/2003/04/emma"
    xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
    xsi:schemaLocation="http://www.w3.org/2003/04/emma
     http://www.w3.org/TR/2007/WD-emma-20070409/emma.xsd"
    xmlns="http://www.example.com/example">
  <emma:interpretation id="int1"
      emma:tokens="From Cambridge to London tomorrow">
    <origin emma:tokens="From Cambridge">Cambridge</origin>
    <destination emma:tokens="to London">London</destination>
    <date emma:tokens="tomorrow">20030315</date>
  </emma:interpretation>
</emma:emma>

4.2.2 Reference to processing: emma:process attribute

Annotation emma:process
Definition An attribute of type xsd:anyURI referencing the process used to generate the interpretation.
Applies to emma:interpretation, emma:one-of, emma:group, emma:sequence

A reference to the information concerning the processing that was used for generating an interpretation MAY be made using the emma:process attribute. For example:

<emma:emma version="1.0"
    xmlns:emma="http://www.w3.org/2003/04/emma"
    xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
    xsi:schemaLocation="http://www.w3.org/2003/04/emma
     http://www.w3.org/TR/2007/WD-emma-20070409/emma.xsd"
    xmlns="http://www.example.com/example">
  <emma:derivation>
    <emma:interpretation id="raw">
      <answer>From Boston to Denver tomorrow</answer>
    </emma:interpretation>

    <emma:interpretation id="better"
        emma:process="http://example.com/mysemproc1.xml">
      <origin>Boston</origin>
      <destination>Denver</destination>
      <date>tomorrow</date>
      <emma:derived-from resource="#raw"/>
    </emma:interpretation>
  </emma:derivation>

  <emma:interpretation id="best"
      emma:process="http://example.com/mysemproc2.xml">
    <origin>Boston</origin>
    <destination>Denver</destination>
    <date>03152003</date>
    <emma:derived-from resource="#better"/>
  </emma:interpretation>
</emma:emma>

The process description document, referenced by the emma:process annotation MAY include information on the process itself, such as grammar, type of parser, etc. EMMA is not normative about the format of the process description document.

4.2.3 Lack of input: emma:no-input attribute

Annotation emma:no-input
Definition Attribute holding xsd:boolean value that is true if there was no input.
Applies to emma:interpretation

The case of lack of input MUST be annotated as follows:

<emma:emma version="1.0"
    xmlns:emma="http://www.w3.org/2003/04/emma"
    xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
    xsi:schemaLocation="http://www.w3.org/2003/04/emma
     http://www.w3.org/TR/2007/WD-emma-20070409/emma.xsd"
    xmlns="http://www.example.com/example">
  <emma:interpretation id="int1" emma:no-input="true"/>
</emma:emma>

If the emma:interpretation is annotated with emma:no-input="true" then the emma:interpretation MUST be empty.

4.2.4 Uninterpreted input: emma:uninterpreted attribute

Annotation emma:uninterpreted
Definition Attribute holding xsd:boolean value that is true if no interpretation was produced in response to the input
Applies to emma:interpretation

An emma:interpretation element representing input for which no interpretation was produced MUST be annotated with emma:uninterpreted="true". For example:

<emma:emma version="1.0"
    xmlns:emma="http://www.w3.org/2003/04/emma"
    xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
    xsi:schemaLocation="http://www.w3.org/2003/04/emma
    http://www.w3.org/TR/2007/WD-emma-20070409/emma.xsd"
    xmlns="http://www.example.com/example">
  <emma:interpretation id="interp1" emma:uninterpreted="true"/>
</emma:emma>

The notation for uninterpreted input MAY refer to any possible stage of interpretation processing, including raw transcriptions. For instance, no interpretation would be produced for stages performing pure signal capture such as audio recordings. Likewise, if a spoken input was recognized but cannot be parsed by a language understanding component, it can be tagged as emma:uninterpreted as in the following example:

<emma:emma version="1.0"
    xmlns:emma="http://www.w3.org/2003/04/emma"
    xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
    xsi:schemaLocation="http://www.w3.org/2003/04/emma
     http://www.w3.org/TR/2007/WD-emma-20070409/emma.xsd"
    xmlns="http://www.example.com/example">
  <emma:interpretation id="understanding"
      emma:process="http://example.com/mynlu.xml"
      emma:uninterpreted="true"
      emma:tokens="From Cambridge to London tomorrow"/>
</emma:emma>

The emma:interpretation MUST be empty if the emma:interpretation element is annotated with emma:uninterpreted="true".

4.2.5 Human language of input: emma:lang attribute

Annotation emma:lang
Definition An attribute of type xsd:language indicating the language for the input.
Applies to emma:interpretation, emma:group, emma:one-of, emma:sequence, and application instance data.

The emma:lang annotation is used to indicate the human language for the input that it annotates. The values of the emma:lang attribute are language identifiers as defined by IETF Best Current Practice 47 [BCP47]. For example, emma:lang="fr" denotes French, and emma:lang="en-US" denotes US English. emma:lang MAY be applied to any emma:interpretation element. Its annotative scope follows the annotative scope of these elements. Unlike the xml:lang attribute in XML, emma:lang does not specify the language used by element contents or attribute values.

In contrast, the attribute xml:lang in XML 1.0 is used to specify the language used in the contents and attribute values of any element in an XML document. The attribute emma:lang MUST be used if the xml:lang can no longer apply. For example, the contents and attribute values of an element in the EMMA document are from different languages, such as in the case where the input language is in French, and the language of the annotated attributes is in English.

The following example shows the use of emma:lang for annotating an input interpretation.

<emma:emma version="1.0"
    xmlns:emma="http://www.w3.org/2003/04/emma"
    xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
    xsi:schemaLocation="http://www.w3.org/2003/04/emma
     http://www.w3.org/TR/2007/WD-emma-20070409/emma.xsd"
    xmlns="http://www.example.com/example">
  <emma:interpretation id="int1" emma:lang="fr">
    <answer>arretez</answer>
  </emma:interpretation>
</emma:emma>

In order handle inputs involving multiple languages, such as through code switching, the emma:lang tag MAY contain several language identifiers separated by spaces.

<emma:emma version="1.0"
    xmlns:emma="http://www.w3.org/2003/04/emma"
    xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
    xsi:schemaLocation="http://www.w3.org/2003/04/emma
     http://www.w3.org/TR/2007/WD-emma-20070409/emma.xsd"
    xmlns="http://www.example.com/example">
  <emma:interpretation id="int1"
      emma:tokens="please stop arretez s'il vous plait"
      emma:lang="en fr">
    <command> CANCEL </command>
  </emma:interpretation>
</emma:emma>

4.2.6 Reference to signal: emma:signal and emma:signal-size attributes

Annotation emma:signal
Definition An attribute of type xsd:anyURI referencing the input signal.
Applies to emma:interpretation, emma:one-of, emma:group, emma:sequence, and application instance data.
Annotation emma:signal-size
Definition An attribute of type xsd:nonNegativeInteger specifying the size in eight bit octets of the referenced source.
Applies to emma:interpretation, emma:one-of, emma:group, emma:sequence, and application instance data.

A URI reference to the signal that originated the input recognition process MAY be represented in EMMA using the emma:signal annotation.

Here is an example where the reference to a speech signal is represented using the emma:signal annotation on the emma:interpretation element:

<emma:emma version="1.0"
    xmlns:emma="http://www.w3.org/2003/04/emma"
    xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
    xsi:schemaLocation="http://www.w3.org/2003/04/emma
     http://www.w3.org/TR/2007/WD-emma-20070409/emma.xsd"
    xmlns="http://www.example.com/example">
  <emma:interpretation id="intp1"
      emma:signal="http://example.com/signals/sg23.bin">
    <origin>Boston</origin>
    <destination>Denver</destination>
    <date>03152003</date>
  </emma:interpretation>
</emma:emma>

The emma:signal-size annotation can be used to declare the exact size of the associated signal in 8-bit octets. An example of the use of an EMMA document to represent a recording, with emma:signal-size indicating the size is as follows:


<emma:emma version="1.0"
    xmlns:emma="http://www.w3.org/2003/04/emma"
    xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
    xsi:schemaLocation="http://www.w3.org/2003/04/emma
     http://www.w3.org/TR/2007/WD-emma-20070409/emma.xsd"
    xmlns="http://www.example.com/example">
  <emma:interpretation id="intp1"
      emma:medium="acoustic"
      emma:mode="voice"
      emma:function="recording"
      emma:uninterpreted="true"
      emma:signal="http://example.com/signals/recording.mpg"
      emma:signal-size="82102" 
      emma:duration="10000">
  </emma:interpretation>
</emma:emma>

4.2.7 Media type: emma:media-type attribute

Annotation emma:media-type
Definition An attribute of type xsd:string holding the MIME type associated with the signal's data format.
Applies to emma:interpretation, emma:one-of, emma:group, emma:sequence, emma:endpoint, and application instance data.

The data format of the signal that originated the input MAY be represented in EMMA using the emma:media-type annotation. An initial set of MIME media types is defined by [RFC2046].

Here is an example where the media type for the ETSI ES 202 212 audio codec for Distributed Speech Recognition (DSR) is applied to the emma:interpretation element. The example also specifies an optional sampling rate of 8 kHz and maxptime of 40 milliseconds.

<emma:emma version="1.0"
    xmlns:emma="http://www.w3.org/2003/04/emma"
    xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
    xsi:schemaLocation="http://www.w3.org/2003/04/emma
     http://www.w3.org/TR/2007/WD-emma-20070409/emma.xsd"
    xmlns="http://www.example.com/example">
  <emma:interpretation id="intp1"
        emma:signal="http://example.com/signals/signal.dsr"
        emma:media-type="audio/dsr-es202212; rate:8000; maxptime:40">
    <origin>Boston</origin>
    <destination>Denver</destination>
    <date>03152003</date>
  </emma:interpretation>
</emma:emma>

4.2.8 Confidence scores: emma:confidence attribute

Annotation emma:confidence
Definition An attribute of type xsd:decimal in range 0.0 to 1.0, indicating the processor's confidence in the result.
Applies to emma:interpretation, emma:one-of, emma:group, emma:sequence, and application instance data.

The confidence score in EMMA is used to indicate the quality of the input, and if confidence is annotated on an input it MUST be given as the value of emma:confidence. The confidence score MUST be a number in the range from 0.0 to 1.0 inclusive. A value of 0.0 indicates minimum confidence, and a value of 1.0 indicates maximum confidence. Note that emma:confidence represents not only the confidence of the speech recognizer, but rather the confidence of the whatever processor was responsible for creating the EMMA result, based on whatever evidence it has. For a natural language interpretation, for example, this might include semantic heuristics in addition to speech recognition scores. Moreover, the confidence score values do not have to be interpreted as probabilities. In fact confidence score values are platform-dependent, since their computation is likely to differ between platforms and different EMMA processors. Confidence scores are annotated explicitly in EMMA in order to provide this information to the subsequent processes for multimodal interaction. The example below illustrates how confidence scores are annotated in EMMA.

<emma:emma version="1.0"
    xmlns:emma="http://www.w3.org/2003/04/emma"
    xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
    xsi:schemaLocation="http://www.w3.org/2003/04/emma
     http://www.w3.org/TR/2007/WD-emma-20070409/emma.xsd"
    xmlns="http://www.example.com/example">
  <emma:one-of id="nbest1">
    <emma:interpretation id="meaning1" emma:confidence="0.6">
      <location>Boston</location>
    </emma:interpretation>

    <emma:interpretation id="meaning2" emma:confidence="0.4">
      <location> Austin </location>
    </emma:interpretation>
  </emma:one-of>
</emma:emma>

In addition to its use as an attribute on the EMMA interpretation and container elements, the emma:confidence attribute MAY also be used to assign confidences to elements in instance data in the application namespace. This can be seen in the following example, where the <destination> and <origin> elements have confidences.

<emma:emma version="1.0"
    xmlns:emma="http://www.w3.org/2003/04/emma"
    xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
    xsi:schemaLocation="http://www.w3.org/2003/04/emma
     http://www.w3.org/TR/2007/WD-emma-20070409/emma.xsd"
    xmlns="http://www.example.com/example">
  <emma:interpretation id="meaning1" emma:confidence="0.6">
     <destination emma:confidence="0.8"> Boston</destination>
     <origin emma:confidence="0.6"> Austin </origin>
  </emma:interpretation>
</emma:emma>

Although in general instance data can be represented in XML using a combination of elements and attributes in the application namespace, EMMA does not provide a standard way to annotate processors' confidences in attributes. Consequently, instance data that is expected to be assigned confidences SHOULD be represented using elements, as in the above example.

4.2.9 Input source: emma:source attribute

Annotation emma:source
Definition An attribute of type xsd:anyURI referencing the source of input.
Applies to emma:interpretation, emma:one-of, emma:group , emma:sequence, and application instance data.

The source of an interpreted input MAY be represented in EMMA as a URI resource using the emma:source annotation.

Here is an example that shows different input sources for different input interpretations.

<emma:emma version="1.0"
    xmlns:emma="http://www.w3.org/2003/04/emma"
    xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
    xsi:schemaLocation="http://www.w3.org/2003/04/emma
     http://www.w3.org/TR/2007/WD-emma-20070409/emma.xsd"
    xmlns="http://www.example.com/example"
    xmlns:myapp="http://www.example.com/myapp">
  <emma:one-of id="nbest1">
    <emma:interpretation id="intp1"
        emma:source="http://example.com/microphone/NC-61">
      <myapp:destination>Boston</myapp:destination>
    </emma:interpretation>

    <emma:interpretation id="intp2"
        emma:source="http://example.com/microphone/NC-4024">
      <myapp:destination>Austin</myapp:destination>
    </emma:interpretation>
  </emma:one-of>
</emma:emma>

4.2.10 Timestamps

The start and end times for input MAY be indicated using either absolute timestamps or relative timestamps. Both are in milliseconds for ease in processing timestamps. Note that the ECMAScript Date object's getTime() function is a convenient way to determine the absolute time.

4.2.10.1 Absolute timestamps: emma:start, emma:end attributes

Annotation emma:start, emma:end
Definition Attributes of type xsd:nonNegativeInteger indicating the absolute starting and ending times of an input in terms of the number of milliseconds since 1 January 1970 00:00:00 GMT
Applies to emma:interpretation, emma:group, emma:one-of, emma:sequence, emma:arc, emma:node, and application instance data

Here is an example of a timestamp for an absolute time.

<emma:emma version="1.0"
    xmlns:emma="http://www.w3.org/2003/04/emma"
    xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
    xsi:schemaLocation="http://www.w3.org/2003/04/emma
     http://www.w3.org/TR/2007/WD-emma-20070409/emma.xsd"
    xmlns="http://www.example.com/example">
  <emma:interpretation id="int1"
       emma:start="1087995961542"
       emma:end="1087995963542">
    <destination>Chicago</destination>
  </emma:interpretation>
</emma:emma>

The emma:start and emma:end annotations on an input MAY be identical, however the emma:end value MUST NOT be less than the emma:start value.

4.2.10.2 Relative timestamps: emma:time-ref-uri, emma:time-ref-anchor-point, emma:offset-to-start attributes

Annotation emma:time-ref-uri
Definition Attribute of type xsd:anyURI indicating the URI used to anchor the relative timestamp.
Applies to emma:interpretation, emma:group, emma:one-of, emma:sequence, emma:lattice, and application instance data
Annotation emma:time-ref-anchor-point
Definition Attribute with a value of start or end, defaulting to start. It indicates whether to measure the time from the start or end of the interval designated with emma:time-ref-uri.
Applies to emma:interpretation, emma:group, emma:one-of, emma:sequence, emma:lattice, and application instance data
Annotation emma:offset-to-start
Definition Attribute of type xsd:integer, defaulting to zero. It specifies the offset in milliseconds for the start of input from the anchor point designated with emma:time-ref-uri and emma:time-ref-anchor-point
Applies to emma:interpretation, emma:group, emma:one-of, emma:sequence, emma:arc, emma:node, and application instance data

Relative timestamps define the start of an input relative to the start or end of a reference interval such as another input.

relative timestamps

The reference interval is designated with emma:time-ref-uri attribute. This MAY be combined with emma:time-ref-anchor-point attribute to specify whether the anchor point is the start or end of this interval. The start of an input relative to this anchor point is then specified with emma:offset-to-start attribute.

Here is an example where the referenced input is in the same document:

<emma:emma version="1.0"
    xmlns:emma="http://www.w3.org/2003/04/emma"
    xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
    xsi:schemaLocation="http://www.w3.org/2003/04/emma
     http://www.w3.org/TR/2007/WD-emma-20070409/emma.xsd"
    xmlns="http://www.example.com/example">
  <emma:sequence>
    <emma:interpretation id="int1">
    <origin>Denver</origin>
    </emma:interpretation>
    <emma:interpretation id="int2"
        emma:time-ref-uri="#int1"
        emma:time-ref-anchor-point="start"
        emma:offset-to-start="5000">
    <destination>Chicago</destination>
    </emma:interpretation>
  </emma:sequence>
</emma:emma>

Note that the reference point refers to an input, but not necessarily to a complete input. For example, if a speech recognizer timestamps each word in an utterance, the anchor point might refer to the timestamp for just one word.

The absolute and relative timestamps are not mutually exclusive; that is, it is possible to have both relative and absolute timestamp attributes on the same EMMA container element.

Timestamps of inputs collected by different devices will be subject to variation if the times maintained by the devices are not synchronized. This concern is outside of the scope of the EMMA specification.

4.2.10.3 Duration of input: emma:duration attribute

Annotation emma:duration
Definition Attribute of type xsd:nonNegativeInteger, defaulting to zero. It specifies the duration of the input in milliseconds.
Applies to emma:interpretation, emma:group, emma:one-of, emma:sequence, emma:arc, and application instance data

The duration of an input in milliseconds MAY be specified with the emma:duration attribute. The emma:duration attribute MAY be used either in combination with timestamps or independently, for example in the annotation of speech corpora.

In the following example, the duration of the signal that gave rise to the interpretation is indicated using emma:duration.

<emma:emma version="1.0"
    xmlns:emma="http://www.w3.org/2003/04/emma"
    xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
    xsi:schemaLocation="http://www.w3.org/2003/04/emma
     http://www.w3.org/TR/2007/WD-emma-20070409/emma.xsd"
    xmlns="http://www.example.com/example">
    <emma:interpretation id="int1" emma:duration="2300">
    <origin>Denver</origin>
    </emma:interpretation>
</emma:emma>

4.2.10.4 Composite Input and Relative Timestamps

This section is informative.

The following table provides guidance on how to determine the values of relative timestamps on a composite input.

Informative Guidance on Relative Timestamps in Composite Derivations
emma:time-ref-uri If the reference interval URI is the same for both inputs then it should be the same for the composite input. If it is not the same then relative timestamps will have to be resolved to absolute timestamps in order to determine the combined timestamp. .
emma:time-ref-anchor-point If the anchor value is the same for both inputs then it should be the same for the composite input. If it is not the same then relative timestamps will have to be resolved to absolute timestamps in order to determine the combined timestamp.
emma:offset-to-start Given that the emma:time-ref-uri and emma:time-ref-anchor-point are the same for both combining inputs, then the emma:offset-to-start for the combination should be the lesser of the two. If they are not the same then relative timestamps will have to be resolved to absolute timestamps in order to determine the combined timestamp.
emma:duration Given that the emma:time-ref-uri and emma:time-ref-anchor-point are the same for both combining inputs, then the emma:duration is calculated as follows. Add together the emma:offset-to-start and emma:duration for each of the inputs. Take whichever of these is greater and subtract from it the lesser of the emma:offset-to-start values in order to determine the combined duration. If emma:time-ref-uri and emma:time-ref-anchor-point are not the same then relative timestamps will have to be resolved to absolute timestamps in order to determine the combined timestamp.

4.2.11 Medium, mode, and function of user inputs: emma:medium, emma:mode, emma:function, emma:verbal attributes

Annotation emma:medium
Definition An attribute of type xsd:nmtokens which contains a space delimited set of values from the set {acoustic, tactile, visual}.
Applies to emma:interpretation, emma:group, emma:one-of, emma:sequence, emma:endpoint, and application instance data
Annotation emma:mode
Definition An attribute of type xsd:nmtokens which contains a space delimited set of values from an open set of values including: {voice, dtmf, ink, gui, keys, video, photograph, ...}.
Applies to emma:interpretation, emma:group, emma:one-of, emma:sequence, emma:endpoint, and application instance data
Annotation emma:function
Definition An attribute of type xsd:string constrained to values in the open set {recording, transcription, dialog, verification, ...}.
Applies to emma:interpretation, emma:group, emma:one-of, emma:sequence, and application instance data
Annotation emma:verbal
Definition An attribute of type xsd:boolean.
Applies to emma:interpretation, emma:group, emma:one-of, emma:sequence, and application instance data

EMMA provides two properties for the annotation of input modality. One indicating the broader medium or channel (emma:medium) and another indicating the specific mode of communication used on that channel (emma:mode). The input medium is defined from the users perspective and indicates whether they use their voice (acoustic), touch (tactile), or visual appearance/motion (visual) as input. Tactile includes most hand-on input device types such as pen, mouse, keyboard, and touch screen. Visual is used for camera input.

emma:medium = space delimited sequence of values from the set: 
            [acoustic|tactile|visual]

The mode property provides the ability to distinguish between different modes of communication that may be within a particular medium. For example, in the tactile medium, modes include electronic ink (ink), and pointing and clicking on a graphical user interface (gui).

emma:mode = space delimited sequence of values from the set:  
            [voice|dtmf|ink|gui|keys|video|photograph| ... ]

The emma:medium classification is based on the boundary between the user and the device that they use. For emma:medium="tactile" the user physically touches the device in order to provide input. For emma:medium="visual" the user's movement is captured by sensors (cameras, infrared) resulting in an input to the system. In the case where emma:medium="acoustic" the user provides input to the system by producing an acoustic signal. Note then that DTMF input will be classified as emma:medium="tactile" since in order to provide DTMF input the user physically presses keys on a keypad.

Orthogonal to the mode, user inputs can also be classified with respect to their communicative function. This enables a simpler mode classification.

emma:function = [recording|transcription|dialog|verification| ... ]

For example, speech can be used for recording (e.g. voicemail), transcription (e.g. dictation), dialog (e.g. interactive spoken dialog systems), and verification (e.g. identifying users through their voiceprints).

EMMA also supports an additional property emma:verbal which distinguishes verbal use of an input mode from non-verbal. This MAY be used to distinguish the use of electronic ink to convey handwritten commands from the user of electronic ink for symbolic gestures such as circles and arrows. Handwritten commands, such as writing downtown in order to change a map display to show the downtown are classified as verbal (emma:function="dialog" emma:verbal="true"). Pen gestures (arrows, lines, circles, etc), such as circling a building, are classified as non-verbal dialog (emma:function="dialog" emma:verbal="false"). The use of handwritten words to transcribe an email message is classified as transcription (emma:function="transcription" emma:verbal="true").

emma:verbal = [true|false]

Handwritten words and ink gestures are typically recognized using different kinds of recognition components (handwriting recognizer vs. gesture recognizer) and the verbal annotation will be added by the recognition component which classifies the input. The original input source, a pen in this case, will not be aware of this difference. The input source identifier will tell you that the input was from a pen of some kind but will not tell you if the mode of input was handwriting (show downtown) or gesture (e.g. circling an object or area).

Here is an example of the EMMA annotation for a pen input where the user's ink is recognized as either a word ("Boston") or as an arrow:

<emma:emma version="1.0"
    xmlns:emma="http://www.w3.org/2003/04/emma"
    xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
    xsi:schemaLocation="http://www.w3.org/2003/04/emma
     http://www.w3.org/TR/2007/WD-emma-20070409/emma.xsd"
    xmlns="http://www.example.com/example">
  <emma:one-of id="nbest1">
    <emma:interpretation id="interp1"
     emma:confidence="0.6"
     emma:medium="tactile"
     emma:mode="ink"
     emma:function="dialog"
     emma:verbal="true">
       <location>Boston</location>
    </emma:interpretation>

    <emma:interpretation id="interp2"
     emma:confidence="0.4"
     emma:medium="tactile"
     emma:mode="ink"
     emma:function="dialog"
     emma:verbal="false">
       <direction>45</direction>
    </emma:interpretation>
  </emma:one-of>
</emma:emma>

Here is an example of the EMMA annotation for a spoken command which is recognized as either "Boston" or "Austin":

<emma:emma version="1.0"
    xmlns:emma="http://www.w3.org/2003/04/emma"
    xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
    xsi:schemaLocation="http://www.w3.org/2003/04/emma
     http://www.w3.org/TR/2007/WD-emma-20070409/emma.xsd"
    xmlns="http://www.example.com/example">
  <emma:one-of>
    <emma:interpretation id="interp1"
     emma:confidence="0.6"
     emma:medium="acoustic"
     emma:mode="voice"
     emma:function="dialog"
     emma:verbal="true">
       <location>Boston</location>
    </emma:interpretation>

    <emma:interpretation id="interp2"
     emma:confidence="0.4"
     emma:medium="acoustic"
     emma:mode="voice"
     emma:function="dialog"
     emma:verbal="true">
       <location>Austin</location>
    </emma:interpretation>
  </emma:one-of>
</emma:emma>

The following table shows the relationship between the medium, mode, and function properties and serves as an aid for classifying inputs. For the dialog function it also shows some examples of the classification of inputs as verbal vs. non-verbal.

Medium Device Mode Function
recording dialog transcription verification
acoustic microphone voice audiofile (e.g. voicemail) spoken command / query / response (verbal = true) dictation speaker recognition
singing a note (verbal = false)
tactile keypad dtmf audiofile / character stream typed command / query / response (verbal = true) text entry (T9-tegic, word completion, or word grammar) password / pin entry
command key "Press 9 for sales" (verbal = false)
keyboard dtmf character / key-code stream typed command / query / response (verbal = true) typing password / pin entry
command key "Press S for sales" (verbal = false)
pen ink trace, sketch handwritten command / query / response (verbal = true) handwritten text entry signature, handwriter recognition
gesture (e.g. circling building) (verbal = false)
gui N/A tapping on named button (verbal = true) soft keyboard password / pin entry
drag and drop, tapping on map (verbal = false)
mouse ink trace, sketch handwritten command / query / response (verbal = true) handwritten text entry N/A
gesture (e.g. circling building) (verbal = false)
gui N/A clicking named button (verbal = true) soft keyboard password / pin entry
drag and drop, clicking on map (verbal = false)
joystick ink trace,sketch gesture (e.g. circling building) (verbal = false) N/A N/A
gui N/A pointing, clicking button / menu (verbal = false) soft keyboard password / pin entry
visual page scanner photograph image handwritten command / query / response (verbal = true) optical character recognition, object/scene recognition (markup, e.g. SVG) N/A
drawings and images (verbal = false)
still camera photograph image objects (verbal = false) visual object/scene recognition face id, retinal scan
video camera video movie sign language (verbal = true) audio/visual recognition face id, gait id, retinal scan
face / hand / arm / body gesture (e.g. pointing, facing) (verbal = false)

4.2.12 Composite multimodality: emma:hook attribute

Annotation emma:hook
Definition An attribute of type xsd:string constrained to values in the open set {voice, dtmf, ink, gui, keys, video, photograph, ...} or the wildcard any
Applies to Application instance data

The attribute emma:hook MAY be used to mark the elements in the application semantics within an emma:interpretation which are expected to be integrated with content from input in another mode to yield a complete interpretation. The emma:mode to be integrated at that point in the application semantics is indicated as the value of the emma:hook attribute. The possible values of emma:hook are the list of input modes that can be values of emma:mode (see Section 4.2.11). In addition to these, the value of emma:hook can also be the wildcard any indicating that the other content can come from any source. The annotation emma:hook differs in semantics from emma:mode as follows. Annotating an element in the application semantics with emma:mode="ink" indicates that that part of the semantics came from the ink mode. Annotating an element in the application semantics with emma:hook="ink" indicates that part of the semantics needs to be integrated with content from the ink mode.

To illustrate the use of emma:hook consider an example composite input in which the user says "zoom in here" in the speech input mode while drawing an area on a graphical display in the ink input mode. The fact that the location element needs to come from the ink mode is indicated by annotating this application namespace element using emma:hook

<emma:emma version="1.0"
    xmlns:emma="http://www.w3.org/2003/04/emma"
    xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
    xsi:schemaLocation="http://www.w3.org/2003/04/emma
     http://www.w3.org/TR/2007/WD-emma-20070409/emma.xsd"
    xmlns="http://www.example.com/example">
  <emma:interpretation emma:mode="voice">
    <command>
      <action>zoom</action>
      <location emma:hook="ink">
        <type>area</type>
      </location>
    </command>
  </emma:interpretation>
</emma:emma>

For more detailed explanation of this example see Appendix C.

4.2.13 Cost: emma:cost attribute

Annotation emma:cost
Definition An attribute of type xsd:decimal in range 0.0 to 10000000, indicating the processor's cost or weight associated with an input or part of an input.
Applies to emma:interpretation, emma:group, emma:one-of, emma:sequence, emma:arc, emma:node, and application instance data.

The cost annotation in EMMA indicates the weight or cost associated with an user's input or part of their input. The most common use of emma:cost is for representing the costs encoded on a lattice output from speech recognition or other recognition or understanding processes. emma:cost MAY also be used to indicate the total cost associated with particular recognition results or semantic intepretations.

<emma:emma version="1.0"
    xmlns:emma="http://www.w3.org/2003/04/emma"
    xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
    xsi:schemaLocation="http://www.w3.org/2003/04/emma
     http://www.w3.org/TR/2007/WD-emma-20070409/emma.xsd"
    xmlns="http://www.example.com/example">
  <emma:one-of>
    <emma:interpretation id="meaning1" emma:cost="1600">
      <location>Boston</location>
    </emma:interpretation>

    <emma:interpretation id="meaning2" emma:cost="400">
      <location> Austin </location>
    </emma:interpretation>
  </emma:one-of>
</emma:emma>

4.2.14 Endpoint properties: emma:endpoint-role, emma:endpoint-address, emma:port-type, emma:port-num, emma:message-id, emma:service-name, emma:endpoint-pair-ref attributes

Annotation emma:endpoint-role
Definition An attribute of type xsd:string constrained to values in the set {source, sink, reply-to, router}.
Applies to emma:endpoint
Annotation emma:endpoint-address
Definition An attribute of type xsd:anyURI that uniquely specifies the network address of the emma:endpoint.
Applies to emma:endpoint
Annotation emma:port-type
Definition An attribute of type xsd:QName that specifies the type of the port.
Applies to emma:endpoint
Annotation emma:port-num
Definition An attribute of type xsd:nonNegativeInteger that specifies the port number.
Applies to emma:endpoint
Annotation emma:message-id
Definition An attribute of type xsd:anyURI that specifies the message ID associated with the data.
Applies to emma:endpoint
Annotation emma:service-name
Definition An attribute of type xsd:string that specifies the name of the service.
Applies to emma:endpoint
Annotation emma:endpoint-pair-ref
Definition An attribute of type xsd:anyURI that specifies the pairing between sink and source endpoints.
Applies to emma:endpoint
Annotation emma:endpoint-info-ref
Definition An attribute of type xsd:IDREF referring to the id attribute of an emma:endpoint-info element.
Applies to emma:interpretation, emma:group, emma:one-of, emma:sequence, and application instance data.

The emma:endpoint-role attribute specifies the role that the particular emma:endpoint performs in multimodal interaction. The role value sink indicates that the particular endpoint is the receiver of the input data. The role value source indicates that the particular endpoint is the sender of the input data. The role value reply-to indicates that the particular emma:endpoint is the intended endpoint for the reply. The same emma:endpoint-address MAY appear in multiple emma:endpoint elements, provided that the same endpoint address is used to serve multiple roles, e.g. sink, source, reply-to, router, etc., or associated with multiple interpretations.

The emma:endpoint-address specifies the network address of the emma:endpoint, and emma:port-type specifies the port type of the emma:endpoint. The emma:port-num annotates the port number of the endpoint (e.g. the typical port number for an http endpoint is 80). The emma:message-id annotates the message ID information associated with the annotated input. This meta information is used to establish and maintain the communication context for both inbound processing and outbound operation. The service specification of the emma:endpoint is annotated by emma:service-name which contains the definition of the service that the emma:endpoint performs. The matching of the sink endpoint and its pairing source endpoint is annotated by the emma:endpoint-pair-ref attribute. One sink endpoint MAY link to multiple source endpoints through emma:endpoint-pair-ref. Further bounding of the emma:endpoint is possible by using the annotation of emma:group (see Section 3.3.2).

The emma:endpoint-info-ref attribute associates the EMMA result in the container element with an emma:endpoint-info element.

The following example illustrates the use of these attributes in multimodal interactions where multiple modalities are used.

<emma:emma version="1.0"
    xmlns:emma="http://www.w3.org/2003/04/emma"
    xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
    xsi:schemaLocation="http://www.w3.org/2003/04/emma
     http://www.w3.org/TR/2007/WD-emma-20070409/emma.xsd"
    xmlns="http://www.example.com/example"
    xmlns: ex="http://www.example.com/emma/port">
  <emma:endpoint-info id="audio-channel-1" >
    <emma:endpoint id="endpoint1"
        emma:endpoint-role="sink"
        emma:endpoint-address="135.61.71.103"
        emma:port-num="50204"
        emma:port-type="rtp"
        emma:endpoint-pair-ref="endpoint2"
        emma:media-type="audio/dsr-202212; rate:8000; maxptime:40"
        emma:service-name="travel"
        emma:mode="voice">
      <ex:app-protocol>SIP</ex:app-protocol>
    </emma:endpoint>

    <emma:endpoint id="endpoint2" emma:endpoint-role="source"
        emma:endpoint-address="136.62.72.104"
        emma:port-num="50204"
        emma:port-type="rtp"
        emma:endpoint-pair-ref="endpoint1"
        emma:media-type="audio/dsr-202212; rate:8000; maxptime:40"
        emma:service-name="travel"
        emma:mode="voice">
      <ex:app-protocol>SIP</ex:app-protocol>
    </emma:endpoint>
  </emma:endpoint-info>

  <emma:endpoint-info id="ink-channel-1">
     <emma:endpoint id="endpoint3" emma:endpoint-role="sink"
         emma:endpoint-address="http://emma.example/sink"
         emma:endpoint-pair-ref="endpoint4"
         emma:port-num="80" emma:port-type="http"
         emma:message-id="uuid:2e5678"
         emma:service-name="travel"
         emma:mode="ink"/>
     <emma:endpoint id="endpoint4"
         emma:endpoint-role="source"
         emma:port-address="http://emma.example/source"
         emma:endpoint-pair-ref="endpoint3"
         emma:port-num="80"
         emma:port-type="http"
         emma:message-id="uuid:2e5678"
         emma:service-name="travel"
         emma:mode="ink"/>
  </emma:endpoint-info>

  <emma:group>
    <emma:interpretation id="int1" emma:start="1087995961542"
        emma:end="1087995963542"
        emma:endpoint-info-ref="audio-channel-1">
      <destination>Chicago</destination>
    </emma:interpretation>

    <emma:interpretation id="int2" emma:start="1087995961542"
        emma:end="1087995963542"
        emma:endpoint-info-ref="ink-channel-1">
      <location>
         <type>area</type>
         <points>34.13 -37.12 42.13 -37.12 ... </points>
      </location>
    </emma:interpretation>
  </emma:group>
</emma:emma>

4.2.15 Reference to emma:grammar element: emma:grammar-ref attribute

Annotation emma:grammar-ref
Definition An attribute of type xsd:IDREF referring to the id attribute of an emma:grammar element.
Applies to emma:interpretation, emma:group, emma:one-of, emma:sequence.

The emma:grammar-ref annotation associates the EMMA result in the container element with an emma:grammar element.

Example:

<emma:emma version="1.0"
    xmlns:emma="http://www.w3.org/2003/04/emma"
    xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
    xsi:schemaLocation="http://www.w3.org/2003/04/emma
     http://www.w3.org/TR/2007/WD-emma-20070409/emma.xsd"
    xmlns="http://www.example.com/example">
  <emma:grammar id="gram1" ref="someURI"/>

  <emma:grammar id="gram2" ref="anotherURI"/>

  <emma:one-of id="r1">
    <emma:interpretation id="int1" emma:grammar-ref="gram1">
      <origin>Boston</origin>
    </emma:interpretation>

    <emma:interpretation id="int2" emma:grammar-ref="gram1">
      <origin>Austin</origin>
    </emma:interpretation>

    <emma:interpretation id="int3" emma:grammar-ref="gram2">
      <command>help</command>
    </emma:interpretation>
  </emma:one-of>
</emma:emma>

4.2.16 Reference to emma:model element: emma:model-ref attribute

Annotation emma:model-ref
Definition An attribute of type xsd:IDREF referring to the id attribute of an emma:model element.
Applies to emma:interpretation, emma:group, emma:one-of, emma:sequence, and application instance data.

The emma:model-ref annotation associates the EMMA result in the container element with an emma:model element.

Example:

<emma:emma version="1.0"
    xmlns:emma="http://www.w3.org/2003/04/emma"
    xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
    xsi:schemaLocation="http://www.w3.org/2003/04/emma
     http://www.w3.org/TR/2007/WD-emma-20070409/emma.xsd"
    xmlns="http://www.example.com/example">
  <emma:model id="model1" ref="someURI"/>

  <emma:model id="model2" ref="anotherURI"/>

  <emma:one-of id="r1">
    <emma:interpretation id="int1" emma:model-ref="model1">
      <origin>Boston</origin>
    </emma:interpretation>

    <emma:interpretation id="int2" emma:model-ref="model1">
      <origin>Austin</origin>
    </emma:interpretation>

    <emma:interpretation id="int3" emma:model-ref="model2">
      <command>help</command>
    </emma:interpretation>
  </emma:one-of>
</emma:emma>

4.2.17 Dialog turns: emma:dialog-turn attribute

Annotation emma:dialog-turn
Definition An attribute of type xsd:string referring to the dialog turn associated with a given container element.
Applies to emma:interpretation, emma:group, emma:one-of, and emma:sequence.

The emma:dialog-turn annotation associates the EMMA result in the container element with a dialog turn. The syntax and semantics of dialog turns is left open to suit the needs of individual applications. For example, some applications might use an integer value, where successive turns are represented by successive integers. Other applications might combine a name of a dialog participant with an integer value representing the turn number for that participant. Ordering semantics for comparison of emma:dialog-turn is deliberately unspecified and left for applications to define.

Example:


<emma:emma version="1.0"
    emma="http://www.w3.org/2003/04/emma"
    xsi="http://www.w3.org/2001/XMLSchema-instance"
    xsi:schemaLocation="http://www.w3.org/2003/04/emma
     http://www.w3.org/TR/2007/WD-emma-20070409/emma.xsd"
    xmlns="http://www.example.com/example">
  <emma:interpretation id="int1" emma:dialog-turn="u8">
    <quantity>3</quantity>
  </emma:interpretation>
</emma:emma>

4.3 Scope of EMMA annotations

The emma:derived-from element (Section 4.1.2) can be used to capture both sequential and composite derivations. This section concerns the scope of EMMA annotations across sequential derivations of user input connected using the emma:derived-from element (Section 4.1.2). Sequential derivations involve processing steps that do not involve multimodal integration, such as applying natural language understanding and then reference resolution to a speech transcription. EMMA derivations describe only single turns of user input and are not intended to describe a sequence of dialog turns.

For example, an EMMA document could contain emma:interpretation elements for the transcription, interpretation, and reference resolution of a speech input, utilizing the id values: raw, better, and best respectively:

<emma:emma version="1.0"
    xmlns:emma="http://www.w3.org/2003/04/emma"
    xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
    xsi:schemaLocation="http://www.w3.org/2003/04/emma
     http://www.w3.org/TR/2007/WD-emma-20070409/emma.xsd"
    xmlns="http://www.example.com/example">
 <emma:derivation>
  <emma:interpretation id="raw"
      emma:process="http://example.com/myasr1.xml"/>
    <answer>From Boston to Denver tomorrow</answer>
  </emma:interpretation>

  <emma:interpretation id="better"
      emma:process="http://example.com/mynlu1.xml">
    <emma:derived-from resource="#raw" composite="false"/>
    <origin>Boston</origin>
    <destination>Denver</destination>
    <date>tomorrow</date>
  </emma:interpretation>
 </emma:derivation>

  <emma:interpretation id="best"
      emma:process="http://example.com/myrefresolution1.xml">
    <emma:derived-from resource="#better" composite="false"/>
    <origin>Boston</origin>
    <destination>Denver</destination>
    <date>03152003</date>
  </emma:interpretation>
</emma:emma>

Each member of the derivation chain is linked to the previous one by a derived-from element (Section 4.1.2), which has an attribute resource that provides a pointer to the emma:interpretation from which it is derived. The emma:process annotation (Section 4.2.2) provides a pointer to the process used for each stage of the derivation.

The following EMMA example represents the same derivation as above but with a more fully specified set of annotations:

<emma:emma version="1.0"
    xmlns:emma="http://www.w3.org/2003/04/emma"
    xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
    xsi:schemaLocation="http://www.w3.org/2003/04/emma
     http://www.w3.org/TR/2007/WD-emma-20070409/emma.xsd"
    xmlns="http://www.example.com/example">
  <emma:derivation>
    <emma:interpretation id="raw"
        emma:process="http://example.com/myasr1.xml"
        emma:source="http://example.com/microphone/NC-61"
        emma:signal="http://example.com/signals/sg23.wav"
        emma:confidence="0.6"
        emma:medium="acoustic"
        emma:mode="voice"
        emma:function="dialog"
        emma:verbal="true"
        emma:tokens="from boston to denver tomorrow"
        emma:lang="en-US">
      <answer>From Boston to Denver tomorrow</answer>
    </emma:interpretation>

    <emma:interpretation id="better"
        emma:process="http://example.com/mynlu1.xml"
        emma:source="http://example.com/microphone/NC-61"
        emma:signal="http://example.com/signals/sg23.wav"
        emma:confidence="0.8"
        emma:medium="acoustic"
        emma:mode="voice"
        emma:function="dialog"
        emma:verbal="true"
        emma:tokens="from boston to denver tomorrow"
        emma:lang="en-US">
      <emma:derived-from resource="#raw" composite="false"/>
      <origin>Boston</origin>
      <destination>Denver</destination>
      <date>tomorrow</date>
    </emma:interpretation>
  </emma:derivation>

  <emma:interpretation id="best"
      emma:process="http://example.com/myrefresolution1.xml"
      emma:source="http://example.com/microphone/NC-61"
      emma:signal="http://example.com/signals/sg23.wav"
      emma:confidence="0.8"
      emma:medium="acoustic"
      emma:mode="voice"
      emma:function="dialog"
      emma:verbal="true"
      emma:tokens="from boston to denver tomorrow"
      emma:lang="en-US">
    <emma:derived-from resource="#better" composite="false"/>
    <origin>Boston</origin>
    <destination>Denver</destination>
    <date>03152003</date>
  </emma:interpretation>
</emma:emma>

EMMA annotations on earlier stages of the derivation often remain accurate at later stages of the derivation. Although this can be captured in EMMA by repeating the annotations on each emma:interpretation within the derivation, as in the example above, there are two disadvantages of this approach to annotation. First, the repetition of annotations makes the resulting EMMA documents significantly more verbose. Second, EMMA processors used for intermediate tasks such as natural language understanding and reference resolution will need to read in all of the annotations and write them all out again.

EMMA overcomes these problems by assuming that annotations on earlier stages of a derivation automatically apply to later stages of the derivation unless a new value is specified. Later stages of the derivation essentially inherit annotations from earlier stages in the derivation. For example, if there was an emma:source annotation on the transcription (raw) it would also apply to the later stages of the derivation such as the result of natural language understanding (better) or reference resolution (best).

Because of the assumption in EMMA that annotations have scope over later stages of a sequential derivation, the example EMMA document above can be equivalently represented as follows:

<emma:emma version="1.0"
    xmlns:emma="http://www.w3.org/2003/04/emma"
    xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
    xsi:schemaLocation="http://www.w3.org/2003/04/emma
     http://www.w3.org/TR/2007/WD-emma-20070409/emma.xsd"
    xmlns="http://www.example.com/example">
  <emma:derivation>
    <emma:interpretation id="raw"
        emma:process="http://example.com/myasr1.xml"
        emma:source="http://example.com/microphone/NC-61"
        emma:signal="http://example.com/signals/sg23.wav"
        emma:confidence="0.6"
        emma:medium="acoustic"
        emma:mode="voice"
        emma:function="dialog"
        emma:verbal="true"
        emma:tokens="from boston to denver tomorrow"
        emma:lang="en-US">
      <answer>From Boston to Denver tomorrow</answer>
    </emma:interpretation>

    <emma:interpretation id="better"
        emma:process="http://example.com/mynlu1.xml"
        emma:confidence="0.8">
      <emma:derived-from resource="#raw" composite="false"/>
      <origin>Boston</origin>
      <destination>Denver</destination>
      <date>tomorrow</date>
    </emma:interpretation>
  </emma:derivation>

  <emma:interpretation id="best"
      emma:process="http://example.com/myrefresolution1.xml">
    <emma:derived-from resource="#better" composite="false"/>
    <origin>Boston</origin>
    <destination>Denver</destination>
    <date>03152003</date>
  </emma:interpretation>
</emma:emma>

The fully specified derivation illustrated above is equivalent to the reduced form derivation following it where only annotations with new values are specified at each stage. These two EMMA documents MUST yield the same result when processed by an EMMA processor.

The emma:confidence annotation is respecified on the better interpretation. This indicates the confidence score for natural language understanding, whereas emma:confidence on the raw interpretation indicates the speech recognition confidence score.

In order to determine the full set of annotations that apply to an emma:interpretation element an EMMA processor or script needs to access the annotations directly on that element and for any that are not specified follow the reference in the resource attribute of the emma:derived-from element to add in annotations from earlier stages of the derivation.

The EMMA annotations break down into three groups with respect to their scope in sequential derivations. One group of annotations always holds true for all members of a sequential derivation. A second group is always respecified on each stage of the derivation. A third group may or may not be respecified.

Scope of Annotations in Sequential Derivations
Classification Annotation
Applies to whole derivation emma:signal
emma:signal-size
emma:dialog-turn
emma:source
emma:medium
emma:mode
emma:function
emma:verbal
emma:lang
emma:tokens
emma:start
emma:end
emma:time-ref-uri
emma:time-ref-anchor-point
emma:offset-to-start
emma:duration
Specified at each stage of derivation emma:derived-from
emma:process
May be respecified emma:confidence
emma:cost
emma:grammar-ref
emma:model-ref
emma:no-input
emma:uninterpreted

One potential problem with this annotation scoping mechanism is that earlier annotations could be lost if earlier stages of a derivation were dropped in order to reduce message size. This problem can be overcome by considering annotation scope at the point where earlier derivation stages are discarded and populating the final interpretation in the derivation with all of the annotations which it could inherit. For example, if the raw and better stages were dropped the resulting EMMA document would be:

<emma:emma version="1.0"
    xmlns:emma="http://www.w3.org/2003/04/emma"
    xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
    xsi:schemaLocation="http://www.w3.org/2003/04/emma
     http://www.w3.org/TR/2007/WD-emma-20070409/emma.xsd"
    xmlns="http://www.example.com/example">
  <emma:interpretation id="best"
      emma:start="1087995961542"
      emma:end="1087995963542"
      emma:process="http://example.com/myrefresolution1.xml"
      emma:source="http://example.com/microphone/NC-61"
      emma:signal="http://example.com/signals/sg23.wav"
      emma:confidence="0.8"
      emma:medium="acoustic"
      emma:mode="voice"
      emma:function="dialog"
      emma:verbal="true"
      emma:tokens="from boston to denver tomorrow"
      emma:lang="en-US">
    <emma:derived-from resource="#better" composite="false"/>
    <origin>Boston</origin>
    <destination>Denver</destination>
    <date>03152003</date>
  </emma:interpretation>
</emma:emma>

Annotations on an emma:one-of element are assumed to apply to all of the container elements within the emma:one-of.

If emma:one-of appears with another emma:one-of then annotations on the parent emma:one-of are assumed to apply to the children of the child emma:one-of.

Annotations on emma:group or emma:sequence do not apply to their child elements.

5. Conformance

The contents of this section are normative.

5.1 Conforming EMMA Documents

A document is a Conforming EMMA Document if it meets both the following conditions:

The EMMA specification and these conformance criteria provide no designated size limits on any aspect of EMMA documents. There are no maximum values on the number of elements, the amount of character data, or the number of characters in attribute values.

5.2 Using EMMA with other Namespaces

The EMMA namespace is intended to be used with other XML namespaces as per the Namespaces in XML Recommendation [XMLNS]. Future work by W3C is expected to address ways to specify conformance for documents involving multiple namespaces.

5.3 Conforming EMMA Processors

A EMMA processor is a program that can process and/or generate Conforming EMMA documents.

In a Conforming EMMA Processor, the XML parser MUST be able to parse and process all XML constructs defined by XML 1.1 [XML] and Namespaces in XML [XMLNS]. It is not required that a Conforming EMMA Processor uses a validating XML parser.

A Conforming EMMA Processor MUST correctly understand and apply the semantics of each markup element or attribute as described by this document.

There is, however, no conformance requirement with respect to performance characteristics of the EMMA Processor. For instance, no statement is required regarding the accuracy, speed or other characteristics of output produced by the processor. No statement is made regarding the size of input that a EMMA Processor is required to support.

Appendices

Appendix A. XML and RELAX NG schemata

This section is Normative.

This section defines the formal syntax for EMMA documents in terms of a normative XML Schema.

There are both an XML Schema and RELAX NG Schema for the EMMA markup. The latest version of the XML Schema for EMMA is available at http://www.w3.org/TR/emma/emma.xsd and the RELAX NG Schema can be found at http://www.w3.org/TR/emma/emma.rng.

For stability it is RECOMMENDED that you use the dated URI available at http://www.w3.org/TR/2007/WD-emma-20070409/emma.xsd and http://www.w3.org/TR/2007/WD-emma-20070409/emma.rng.

Appendix B. MIME type

This section is Normative.

This appendix registers a new MIME media type, "application/emma+xml".

B.1 Registration of MIME media type application/emma+xml

MIME media type name:

application

MIME subtype name:

emma+xml

Required parameters:

None.

Optional parameters:
charset

This parameter has identical semantics to the charset parameter of the application/xml media type as specified in [RFC3023] or its successor.

Encoding considerations:

By virtue of EMMA content being XML, it has the same considerations when sent as "application/emma+xml"as does XML. See RFC 3023 (or its successor), section 3.2.

Security considerations:

Several features of EMMA require dereferencing arbitrary URIs. Implementors are adivised to heed the security issues of [RFC3986] section 7.

In addition, because of the extensibility features for EMMA, it is possible that "application/emma+xml" will describe content that has security implications beyond those described here. However, if the processor follows only the normative semantics of this specification, this content will be ignored. Only in the case where the processor recognizes and processes the additional content, or where further processing of that content is dispatched to other processors, would security issues potentially arise. And in that case, they would fall outside the domain of this registration document.

Interoperability considerations:

This specification describes processing semantics that dictate the required behavior for dealing with, among other things, unrecognized elements.

Because EMMA is extensible, conformant "application/emma+xml" processors MAY expect that content received is well-formed XML, but processors SHOULD NOT assume that the content is valid EMMA or expect to recognize all of the elements and attributes in the document.

Published specification:

This media type registration is for EMMA documents as described by this specification.

Additional information:
Magic number(s):

There is no single initial octet sequence that is always present in EMMA documents.

File extension(s):

EMMA documents are most often identified with the extensions ".emma".

Macintosh File Type Code(s):

TEXT

Person & email address to contact for further information:

Kazuyuki Ashimura, <ashimura@w3.org>.

Intended usage:

COMMON

Author/Change controller:

The EMMA specification is a work product of the World Wide Web Consortium's Multimodal Interaction Working Group. The W3C has change control over these specifications.

Appendix C. emma:hook and SRGS

This section is Informative.

One of the most powerful aspects of multimodal interfaces is their ability to provide support for user inputs which are distributed over the available input modes. These composite inputs are contributions made by the user within a single turn which have component parts in different modes. For example, the user might say "zoom in here" in the speech mode while drawing an area on a graphical display in the ink mode. One of the central motivating factors for this kind of input is that different kinds of communicative content are best suited to different input modes. In the example of a user drawing an area on a map and saying "zoom in here", the zoom command is easiest to provide in speech but the spatial information, the specific area, is easier to provide in ink.

Enabling composite multimodality is critical in ensuring that multimodal systems support more natural and effective interaction for users. In order to support composite inputs, a multimodal architecture must provide some kind of multimodal integration mechanism. In the W3C Multimodal Interaction Framework [MMI Framework], multimodal integration can be handled by an integration component which follows the application of speech understanding and other kinds of interpretation procedures for individual modes.

Given the broad range of different techniques being employed for multimodal integration and the extent to which this is an ongoing research problem, standardization of the specific method or algorithm used for multimodal integration is not appropriate at this time. In order to facilitate the development and inter-operation of different multimodal integration mechanisms EMMA provides markup language enabling application independent specification of elements in the application markup where content from another mode needs to be integrated. These representation 'hooks' can then be used by different kinds of multimodal integration components and algorithms to drive the process of multimodal integration. In the processing of a composite multimodal input, the result of applying a mode-specific interpretation component to each of the individual modes will be EMMA markup describing the possible interpretation of that input.

One way to build an EMMA representation of a spoken input such as "zoom in here" is to use grammar rules in the W3C Speech Recognition Grammar Specification [SRGS] using the Semantic Interpretation [SISR] tags to build the application semantics with the emma:hook attribute. In this approach [ECMAScript] is specified in order to build up an object representing the semantics. The resulting ECMAScript object is then translated to XML.

For our example case of "zoom in here". The following SRGS rule could be used. The Semantic Interpretation for Speech Recognition specification [SISR] provides a reserved property _nsprefix for indicating the namespace to be used with an attribute.

<rule id="zoom">
  zoom in here
  <tag>
    $.command = new Object();
    $.command.action = "zoom";
    $.command.location = new Object();
    $.command.location._attributes = new Object();
    $.command.location._attributes.hook = new Object();
    $.command.location._attributes.hook._nsprefix = "emma";
    $.command.location._attributes.hook._value = "ink";
    $.command.location.type = "area";
  </tag>
</rule>

Application of this rule will result in the following ECMAScript object being built.

command: {
      action: "zoom"
      location: {
        _attributes: {
           hook: {
             _nsprefix: "emma"
             _value: "ink"
             }
           }
        type: "area"
      }
}

SI processing in an XML environment would generate the following document:

<command>
  <action>zoom</action>
  <location emma:hook="ink">
     <type>area</type>
  </location>
</command>

This XML fragment might then appear within an EMMA document as follows:

<emma:emma version="1.0"
    xmlns:emma="http://www.w3.org/2003/04/emma"
    xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
    xsi:schemaLocation="http://www.w3.org/2003/04/emma
     http://www.w3.org/TR/2007/WD-emma-20070409/emma.xsd"
    xmlns="http://www.example.com/example">
  <emma:interpretation id="voice1"
      emma:medium="acoustic"
      emma:mode="voice">
    <command>
      <action>zoom</action>
      <location emma:hook="ink">
         <type>area</type>
      </location>
    </command>
  </emma:interpretation>
</emma:emma>

The emma:hook annotation indicates that this speech input needs to be combined with ink input such as the following:

<emma:emma version="1.0"
    xmlns:emma="http://www.w3.org/2003/04/emma"
    xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
    xsi:schemaLocation="http://www.w3.org/2003/04/emma
     http://www.w3.org/TR/2007/WD-emma-20070409/emma.xsd"
    xmlns="http://www.example.com/example">
  <emma:interpretation id="pen1"
      emma:medium="tactile"
      emma:mode="ink">
    <location>
      <type>area</type>
      <points>42.1345 -37.128 42.1346 -37.120 ... </points>
    </location>
  </emma:interpretation>
</emma:emma>

This representation could be generated by a pen modality component performing gesture recognition and interpretation. The input to the component would be an Ink Markup Language specification [INKML] of the ink trace and the output would be the EMMA document above.

The combination will result in the following EMMA document for the combined speech and pen multimodal input.

<emma:emma version="1.0"
    xmlns:emma="http://www.w3.org/2003/04/emma"
    xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
    xsi:schemaLocation="http://www.w3.org/2003/04/emma
     http://www.w3.org/TR/2007/WD-emma-20070409/emma.xsd"
    xmlns="http://www.example.com/example">
  <emma:interpretation
      emma:medium="acoustic tactile" 
      emma:mode="voice ink"
      emma:process="http://example.com/myintegrator.xml">
    <emma:derived-from resource="http://example.com/voice1.emma/#voice1" composite="true"/>
    <emma:derived-from resource="http://example.com/pen1.emma/#pen1" composite="true"/>
    <command>
       <action>zoom</action>
       <location>
         <type>area</type>
         <points>42.1345 -37.128 42.1346 -37.120 ... </points>
        </location>
     </command>
  </emma:interpretation>
</emma:emma>

There are two components to the process of integrating these two pieces of semantic markup. The first is to ensure that the two are compatible; that is, that no semantic constraints are violated. The second is to fuse the content from the two sources. In our example, the <type>area</type> element is intended to indicate that this speech command requires integration with an area gesture rather than, for example, a line gesture, which would have the subelement <type>line</type>. This constraint needs to be enforced by whatever mechanism is responsible for multimodal integration.

Many different techniques could be used for achieving this integration of the semantic interpretation of the pen input, a <location> element, with the corresponding <location> element in the speech. The emma:hook simply serves to indicate the existence of this relationship.

One way to achieve both the compatibility checking and fusion of content from the two modes is to use a well-defined general purpose matching mechanism such as unification. Graph unification [Graph unification] is a mathematical operation defined over directed acylic graphs which captures both of the components of integration in a single operation: the applications of the semantic constraints and the fusing of content. One possible semantics for the emma:hook markup indicates that content from the required mode needs to be unified with that position in the application semantics. In order to unify, two elements must not have any conflicting values for subelements or attributes. This procedure can be defined recursively so that elements within the subelements must also not clash and so on. The result of unification is the union of all of the elements and attributes of the two elements that are being unified.

In addition to the unification operation, in the resulting emma:interpretation the emma:hook attribute needs to be removed and the emma:mode attribute changed to the list of the modes of the individual inputs , e.g. "voice ink".

Instead of the unification operation, for a specific application semantics, integration could be achieved using some other algorithm or script. The benefit of using the unification semantics for emma:hook is that it provides a general purpose mechanism for checking the compatibility of elements and fusing them, whatever the specific elements are in the application specific semantic representation.

The benefit of using the emma:hook annotation for authors is that it provides an application independent method for indicating where integration with content from another mode is required. If a general purpose integration mechanism is used, such as the unification approach described above, authors should be able to use the same integration mechanism for a range of different applications without having to change the integration rules or logic. For each application the speech grammar rules [SRGS] need to assign emma:hook to the appropriate elements in the semantic representation of the speech. The general purpose multimodal integration mechanism will use the emma:hook annotations in order to determine where to add in content from other modes. Another benefit of the emma:hook mechanism is that it facilitates interoperability among different multimodal integration components, so long as they are all general purpose and utilize emma:hook in order to determine where to integrate content.

The following provides a more detailed example of the use of the emma:hook annotation. In this example, spoken input is combined with two ink gestures. The semantic representation assigned to the spoken input "send this file to this" indicates two locations where content is required from ink input using emma:hook="ink":

<emma:emma version="1.0"
    xmlns:emma="http://www.w3.org/2003/04/emma"
    xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
    xsi:schemaLocation="http://www.w3.org/2003/04/emma
     http://www.w3.org/TR/2007/WD-emma-20070409/emma.xsd"
    xmlns="http://www.example.com/example">
  <emma:interpretation id="voice2"
      emma:medium="acoustic"
      emma:mode="voice"
      emma:tokens="send this file to this"
      emma:start="1087995961500"
      emma:end="1087995963542">
    <command>
      <action>send</action>
        <arg1>
          <object emma:hook="ink">
            <type>file</type>
            <number>1</number>
          </object>
        </arg1>
       <arg2>
         <object emma:hook="ink">
           <number>1</number>
         </object>
       </arg2>
    </command>
  </emma:interpretation>
</emma:emma>

The user gesturing on the two locations on the display can be represented using emma:sequence:

<emma:emma version="1.0"
    xmlns:emma="http://www.w3.org/2003/04/emma"
    xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
    xsi:schemaLocation="http://www.w3.org/2003/04/emma
     http://www.w3.org/TR/2007/WD-emma-20070409/emma.xsd"
    xmlns="http://www.example.com/example">
  <emma:sequence id="ink2"
      emma:medium="tactile"
      emma:mode="ink">
    <emma:interpretation emma:start="1087995960500"
      emma:end="1087995960900">
      <object>
       <type>file</type>
       <number>1</number>
       <id>test.pdf</id>
      <object>
    </emma:interpretation>
    <emma:interpretation emma:start="1087995961000"
      emma:end="1087995961100">
      <object>
        <type>printer</type>
        <number>1</number>
        <id>lpt1</id>
      <object>
    </emma:interpretation>
  </emma:sequence>
</emma:emma>

A general purpose unification-based multimodal integration algorithm could use the emma:hook annotation as follows. It identifies the elements marked with emma:hook in document order. For each of those in turn, it attempts to unify the element with the corresponding element in order in the emma:sequence. Since none of the subelements conflict, the unification goes through and as a result, we have the following EMMA for the composite result:

<emma:emma version="1.0"
    xmlns:emma="http://www.w3.org/2003/04/emma"
    xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
    xsi:schemaLocation="http://www.w3.org/2003/04/emma
     http://www.w3.org/TR/2007/WD-emma-20070409/emma.xsd"
    xmlns="http://www.example.com/example">
<emma:interpretation id="multimodal2"
      emma:medium="acoustic tactile"
      emma:mode="voice ink"
      emma:tokens="send this file to this"
      emma:process="http://example.com/myintegration.xml"
      emma:start="1087995960500"
      emma:end="1087995963542">
  <emma:derived-from resource="http://example.com/voice2.emma/#voice2" composite="true"/>
  <emma:derived-from resource="http://example.com/ink2.emma/#ink2" composite="true"/>
  <command>
   <action>send</action>
    <arg1>
     <object>
       <type>file</type>
       <number>1</number>
        <id>test.pdf</id>
     </object>
    </arg1>
    <arg2>
     <object>
       <type>printer</type>
        <number>1</number>
       <id>lpt1</id>
     </object>
    </arg2>
  </command>
</emma:interpretation>
</emma:emma>

Appendix D. EMMA event interface

This section is Informative.

The W3C Document Object Model [DOM] defines platform and language neutral interfaces that gives programs and scripts the means to dynamically access and update the content, structure and style of documents. DOM Events define a generic event system which allows registration of event handlers, describes event flow through a tree structure, and provides basic contextual information for each event.

This section of the EMMA specification extends the DOM Event interface for use with events that describe interpreted user input in terms of a DOM Node for an EMMA document.

// File: emma.idl

#ifndef _EMMA_IDL_
#define _EMMA_IDL_

#include "dom.idl"#include "views.idl"#include "events.idl"
#pragma prefix "dom.w3c.org"module emma
{
  typedef dom::DOMString DOMString;
  typedef dom::Node Node;

  interface EMMAEvent : events::UIEvent {
    readonly attribute dom::Node  node;
    void               initEMMAEvent(in DOMString typeArg,
                                   in boolean canBubbleArg,
                                   in boolean cancelableArg,
                                   in Node node);
  };
};

#endif // _EMMA_IDL_

Appendix E. References

E.1 Normative references

BCP47
A. Phillips and M. Davis, editors. Tags for the Identification of Languages, IETF, September 2006.
RFC3023
M. Murata et al., editors. XML Media Types. IETF RFC 3023, January 2001.
RFC2046
N. Freed and N. Borenstein, editors. Multipurpose Internet Mail Extensions (MIME) Part Two: Media Types. IETF RFC 2046, November 1996.
RFC2119
S. Bradner, editor. Key words for use in RFCs to Indicate Requirement Levels, IETF RFC 2119, March 1997.
RFC3986
T. Berners-Lee et al., editors. Uniform Resource Identifier (URI): Generic Syntax. IETF RFC 3986, January 2005.
RFC3987
M. Duerst and M. Suignard, editors. Internationalized Resource Identifiers (IRIs). IETF RFC 3987, January 2005.
XML
Tim Bray et al., editors. Extensible Markup Language (XML) 1.1. World Wide Web Consortium, W3C Recommendation, 2004.
XMLNS
Tim Bray et al., editors. Namespaces in XML 1.1, World Wide Web Consortium, W3C Recommendation, 2006.
XML Schema Structures
Henry S. Thompson et al., editors. XML Schema Part 1: Structures Second Edition, World Wide Web Consortium, W3C Recommendation, 2004.
XML Schema Datatypes
Paul V. Biron and Ashok Malhotra, editors. XML Schema Part 2: Datatypes Second Edition, World Wide Web Consortium, W3C Recommendation, 2004.

E.2 Informative references

DOM
Document Object Model, World Wide Web Consortium, 2005.
ECMAScript
ECMAScript
INKML
Yi-Min Chee, Max Froumentin, Stephen M. Watt, editors. Ink Markup Language (InkML), World Wide Web Consortium, W3C Working Draft, 2006.
SISR
Luc Van Tichelen and Dave Burke, editors. Semantic Interpretation for Speech Recognition, World Wide Web Consortium, W3C Proposed Recommendation, 2007.
SRGS
Andrew Hunt, Scott McGlashan, editors. Speech Recognition Grammar Specification Version 1.0, World Wide Web Consortium, W3C Recommendation, 2004.
XFORMS
John M. Boyer et al., editors. XForms 1.0 (Second Edition), World Wide Web Consortium, W3C Recommendation, 2006.
RELAX-NG
James Clark and Makoto Murata, editors. RELAX NG Specification, OASIS, Committee Specification, 2001.
EMMA Requirements
Stephane H. Maes and Stephen Potter, editors. Requirements for EMMA, World Wide Web Consortium, W3C Note, 2003.
Graph Unification
Bob Carpenter. The Logic of Typed Feature Structures, Cambridge Tracts in Theoretical Computer Science 32, Cambridge University Press, 1992.
Kevin Knight. Unification: A Multidisciplinary Survey, ACM Computing Surveys, 21(1), 1989.
Michael Johnston. Unification-based Multimodal Parsing, Proceedings of Association for Computational Linguistics, pp. 624-630, 1998.
MMI Framework
James A. Larson, T.V. Raman and Dave Raggett, editors. W3C Multimodal Interaction Framework, World Wide Web Consortium, W3C Note, 2003.
MMI Requirements
Stephane H. Maes and Vijay Saraswat, editors. Multimodal Interaction Requirements, World Wide Web Consortium, W3C Note, 2003.

Appendix F. Changes since last draft

This section is Informative.

Since the publication of the original Last Call working draft, the EMMA specification has undergone numerous editorial changes and corrections and incorporates revisions and clarifications and extensions made in response to detailed feedback from the Internationalization working group, the Web Accessibility Initiative, the Voice Browser working group, and the Smartweb consortium. These changes include marking of normative versus informative material and use of the key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" as described in [RFC2119], to clarify the status of statements in the specification. The references have also been updated.

Appendix G. Acknowledgements

This section is Informative.

The editors would like to recognize the contributions of the current and former members of the W3C Multimodal Interaction Group (listed in alphabetical order). Their current affiliations at time of publication are given:

Kazuyuki Ashimura, W3C
Patrizio Bergallo, Loquendo
Daniel Burnett, Nuance
Wu Chou, Avaya
Max Froumentin, Joost
Katriina Halonen, Nokia
Jin Liu, T-Systems
Roberto Pieraccini, Speechcycle
Stephen Potter, Microsoft
Massimo Romanelli, DFKI
Yuan Shao, Canon