WEBVTT
Kind: captions
Language: en

00:00:00.000 --> 00:00:02.720
Hello, everyone, and welcome back

00:00:03.720 --> 00:00:05.880
to the second panel.

00:00:05.880 --> 00:00:08.800
And I'm now joined by

00:00:10.360 --> 00:00:14.080
Chaohai Ding
from the University of Southampton.

00:00:14.760 --> 00:00:17.840
Lourdes Moreno
from the Universidade

00:00:17.880 --> 00:00:20.880
Carlos III de Madrid
in Spain.

00:00:21.120 --> 00:00:24.120
And Vikas Ashok from

00:00:24.320 --> 00:00:26.960
Old Dominion University in the US.

00:00:27.760 --> 00:00:31.120
So thank you all for your availability.

00:00:31.120 --> 00:00:33.840
It's great to have you here.

00:00:33.840 --> 00:00:37.800
And let's as I said before,

00:00:37.800 --> 00:00:42.000
let's bring back the topic
of natural language processing. We

00:00:43.160 --> 00:00:44.640
we addressed it

00:00:44.640 --> 00:00:49.160
yesterday, but now from the perspective of

00:00:49.160 --> 00:00:54.240
how can it be used to enhance
accessible communication on the Web.

00:00:54.560 --> 00:00:57.680
And so here and I guess

00:00:58.160 --> 00:01:01.120
once again, similar to what I've done

00:01:01.800 --> 00:01:04.720
an hour ago in the first panel,

00:01:04.720 --> 00:01:07.920
you've been working on different aspects
of this

00:01:08.160 --> 00:01:11.400
large domain of accessible communication.

00:01:11.840 --> 00:01:13.440
And you’ve

00:01:13.720 --> 00:01:16.520
pursued advances in machine

00:01:16.520 --> 00:01:20.000
translation, in sign language,

00:01:20.400 --> 00:01:24.400
AAC, and so from your perspective

00:01:24.400 --> 00:01:27.120
and from your focus on the work,

00:01:27.640 --> 00:01:32.160
what are the current challenges
that that you've been facing?

00:01:32.160 --> 00:01:34.200
And that's that are preventing

00:01:35.920 --> 00:01:38.080
the next breakthrough, I guess.

00:01:38.080 --> 00:01:43.560
And also I would like to ask you to,
uh, for your first intervention also to do

00:01:44.080 --> 00:01:47.680
a brief introduction to yourself
and to what you've been doing.

00:01:48.240 --> 00:01:54.080
Okay, So I can start with you
Chaohai.

00:01:54.120 --> 00:01:55.920
Thank you for having me today.

00:01:55.920 --> 00:02:00.720
My name is Chaohai Ding and I'm a senior research
fellow at the University of Southampton.

00:02:01.160 --> 00:02:05.560
And my research interest is
AI and inclusion

00:02:06.200 --> 00:02:09.440
which includes using data science and AI

00:02:10.400 --> 00:02:12.000
techniques to enhance

00:02:12.000 --> 00:02:15.440
accessible learning, traveling and

00:02:16.520 --> 00:02:17.920
communication.

00:02:17.920 --> 00:02:20.600
So, yes, we used...

00:02:21.120 --> 00:02:22.400
NLP has been widely

00:02:22.400 --> 00:02:26.160
used in our research to
support accessible communication.

00:02:27.200 --> 00:02:31.480
Currently, we are working on
several projects focused on AAC.

00:02:31.480 --> 00:02:33.520
So for example,

00:02:33.600 --> 00:02:36.280
we applied the concept

00:02:36.280 --> 00:02:39.960
net of ...  knowledge graph

00:02:40.080 --> 00:02:42.600
to interlinking AAC

00:02:43.920 --> 00:02:46.000
symbols from different symbol sets.

00:02:47.200 --> 00:02:49.120
This can be used

00:02:49.120 --> 00:02:51.800
for symbol to symbol translation.

00:02:52.640 --> 00:02:57.840
And we also developed an NLP model
to translate

00:02:57.840 --> 00:03:02.760
the AAC symbol sequence
into spoke text sequence.

00:03:03.400 --> 00:03:07.200
So so that's the two projects
we're working on currently

00:03:07.920 --> 00:03:12.000
and we also working on 
accessible e-learning project

00:03:12.360 --> 00:03:14.880
that we applied a machine translation

00:03:16.040 --> 00:03:18.480
to provide transcripts

00:03:18.480 --> 00:03:20.600
from English to,

00:03:20.600 --> 00:03:23.400
other languages
for our international users.

00:03:24.000 --> 00:03:27.680
So that's another scenario
we are working with on a machine

00:03:27.680 --> 00:03:30.800
translation for accessible communication.

00:03:30.800 --> 00:03:35.360
So there are a few challenges we have
identified in our kind of research.

00:03:36.240 --> 00:03:38.640
The first one is always the data,

00:03:38.640 --> 00:03:42.640
data availability
and the data opti...bility

00:03:43.160 --> 00:03:48.360
So as we know, an NLP model is normally
trained on a large amount of data. So

00:03:49.360 --> 00:03:52.200
especially for AAC,

00:03:53.720 --> 00:03:57.200
we, we are... 
one of the biggest challenges

00:03:57.200 --> 00:04:00.480
is that we are lack of a

00:04:00.480 --> 00:04:03.360
data like user

00:04:04.240 --> 00:04:07.720
user user data... AAC data,
and also

00:04:08.360 --> 00:04:12.120
how a user interact with the AAC.

00:04:12.520 --> 00:04:14.600
So so

00:04:16.080 --> 00:04:17.600
which...

00:04:17.600 --> 00:04:20.400
and also we have several different

00:04:21.000 --> 00:04:25.800
AAC symbol sets used
by the different individuals

00:04:26.040 --> 00:04:30.600
and which make it very difficult
to develop NLP models as well

00:04:30.840 --> 00:04:33.760
because the AAC symbols

00:04:33.760 --> 00:04:38.040
are separate for each symbol set
and that's the

00:04:38.080 --> 00:04:40.720
another challenge is the lack of data

00:04:41.480 --> 00:04:45.080
interoperability in AAC symbol sets.

00:04:46.320 --> 00:04:51.000
Yet the third challenge
we are identified is the inclusion

00:04:51.240 --> 00:04:54.960
because we are working on AAC symbol sets

00:04:55.720 --> 00:04:59.000
from Arabic, English and Chinese.

00:04:59.280 --> 00:05:05.040
So there are cultural and social
difference in AAC symbols, which is

00:05:06.760 --> 00:05:08.240
important to

00:05:08.240 --> 00:05:13.880
consider the needs of different end
user groups on the cultural and the social

00:05:14.160 --> 00:05:18.240
factors,
and to be involved in the development

00:05:18.360 --> 00:05:20.960
of the NLP models for AAC.

00:05:22.440 --> 00:05:24.160
The first one is

00:05:24.400 --> 00:05:28.120
data privacy and safety 
and this has been

00:05:29.760 --> 00:05:33.360
identified in our web application for

00:05:33.360 --> 00:05:38.760
or from AAC symbols to spoken texts.

00:05:38.760 --> 00:05:42.600
So how do we,
if we want to, a more accurate

00:05:42.600 --> 00:05:44.960
or more specific,

00:05:45.960 --> 00:05:47.400
personalized

00:05:47.960 --> 00:05:48.840
application?

00:05:48.840 --> 00:05:51.000
We need the user's information.

00:05:51.000 --> 00:05:54.000
So the challenge is how

00:05:54.000 --> 00:05:57.160
how can we store this

00:05:57.520 --> 00:05:59.040
personal information

00:05:59.040 --> 00:06:03.000
and how how prevent the data
misuse and the bridge

00:06:03.400 --> 00:06:06.200
and how to make the tradeoff

00:06:06.200 --> 00:06:08.760
between the user

00:06:09.240 --> 00:06:12.000
information and the the model performance.

00:06:13.720 --> 00:06:17.040
And the last one

00:06:17.040 --> 00:06:20.360
is always the accessible user interface.

00:06:20.360 --> 00:06:22.480
So in the how,

00:06:22.480 --> 00:06:26.040
how to makes this AI powered tool, 
NLP powered

00:06:26.280 --> 00:06:30.840
tools accessible
for for end users and

00:06:32.680 --> 00:06:34.040
and also

00:06:34.040 --> 00:06:36.760
there are more generic issues in

00:06:37.760 --> 00:06:42.400
AI like accountability Explainability
so yes I think

00:06:42.400 --> 00:06:46.200
that's the last of the challenges
we have identified in our research.

00:06:47.760 --> 00:06:48.480
Thank you.

00:06:48.720 --> 00:06:49.000
Thank you, Chaohai.

00:06:49.000 --> 00:06:54.920
Is a great summary of definitely 
some of the major challenges

00:06:54.920 --> 00:06:59.000
that well are spread
across the entire domain.

00:06:59.400 --> 00:07:01.200
Definitely. Thank you so much.

00:07:01.200 --> 00:07:04.280
Lourdes, do you want to go next?

00:07:06.400 --> 00:07:11.240
You’re muted.

00:07:11.240 --> 00:07:12.240
Thank you.

00:07:12.840 --> 00:07:14.520
Thanks for the invitation.

00:07:14.520 --> 00:07:16.920
Good afternoon everyone. I’m Lourdes

00:07:16.920 --> 00:07:21.400
Moreno. I work as an associate
professor in the computer

00:07:21.400 --> 00:07:24.880
science department
on the Universidad Carlos III

00:07:24.960 --> 00:07:27.880
de Madrid, Spain.

00:07:27.880 --> 00:07:29.160
I am an accessibility

00:07:29.160 --> 00:07:33.160
expert. I have been working
in the area of technology

00:07:33.160 --> 00:07:38.080
for disability for 20 years. And...

00:07:38.600 --> 00:07:41.200
I have previously worked on

00:07:41.200 --> 00:07:44.120
sensory disability but currently

00:07:44.640 --> 00:07:47.080
I work on cognitive accessibility.

00:07:47.640 --> 00:07:50.560
In my research work I combinate

00:07:50.880 --> 00:07:53.920
method from the human computer interaction

00:07:54.320 --> 00:07:57.720
and natural language processing areas

00:07:58.280 --> 00:08:00.480
to obtain accessible solutions

00:08:00.960 --> 00:08:03.560
from the point of view of reability

00:08:03.560 --> 00:08:07.120
and the stability of the language
in user interfaces.

00:08:08.080 --> 00:08:09.080
And so is

00:08:09.080 --> 00:08:15.320
the question currently in natural
language research is being developed

00:08:15.440 --> 00:08:18.360
at our language model in recent years,

00:08:18.680 --> 00:08:20.960
there had been many advances

00:08:22.080 --> 00:08:24.320
due to the increasing resources

00:08:24.720 --> 00:08:28.680
such as large dataset and cloud platform

00:08:29.240 --> 00:08:32.320
that allow the training of large models.

00:08:33.120 --> 00:08:36.480
But the most crucial factor is the use

00:08:36.480 --> 00:08:41.200
of transforming technology
and the use of transfer learning.

00:08:41.200 --> 00:08:47.640
These are methods based on deep 
learning to create language model

00:08:48.160 --> 00:08:50.280
base of the neural network.

00:08:51.440 --> 00:08:53.400
They are universal models,

00:08:53.400 --> 00:08:57.160
but then support is different
in natural processing

00:08:57.160 --> 00:09:03.240
language tasks, such as 
question answering, translation,

00:09:03.240 --> 00:09:06.920
summarization,
speech recognition and more.

00:09:07.840 --> 00:09:13.840
The most extensively used models
are the GPT

00:09:13.840 --> 00:09:17.280
from OpenAI, 
and Bard from Google.

00:09:17.760 --> 00:09:22.680
But new and bigger models
continually appear

00:09:23.360 --> 00:09:25.720
that outperform previous one

00:09:25.720 --> 00:09:31.200
because they are a performance continuous
to a scale

00:09:31.400 --> 00:09:37.440
as more parameters are added to their
models and more data are added.

00:09:38.640 --> 00:09:42.920
However, and despite these great advance

00:09:43.440 --> 00:09:46.160
there are issues

00:09:46.160 --> 00:09:51.240
in the accessibility scope
challenges to address.

00:09:51.240 --> 00:09:55.280
One of them is bias.

00:09:55.280 --> 00:09:59.000
Language models have
different types of bias

00:09:59.440 --> 00:10:03.240
such as gender, race and disability

00:10:03.840 --> 00:10:07.680
but gender and race

00:10:07.760 --> 00:10:11.040
biases are highly analyzed.

00:10:11.640 --> 00:10:14.640
However it is in the case

00:10:14.720 --> 00:10:16.840
with disability biases.

00:10:17.960 --> 00:10:22.240
It has been a relatively underexplored.

00:10:23.040 --> 00:10:27.600
There are studies relative
this matter for for example

00:10:27.600 --> 00:10:32.880
in in these work
in the in the sentiment and analysis text

00:10:33.400 --> 00:10:37.200
the terms relative to disability

00:10:37.200 --> 00:10:40.320
have a negative value

00:10:40.320 --> 00:10:45.120
or in another work using a
model to moderate conversation

00:10:45.120 --> 00:10:48.840
classified takes mentions of disability

00:10:49.080 --> 00:10:52.360
as more toxics.

00:10:53.200 --> 00:10:57.600
That is algorithms
are trained to be restful

00:10:57.600 --> 00:11:00.040
that can be offensive and cause

00:11:00.040 --> 00:11:03.000
disadvantage to individual
with disabilities.

00:11:04.240 --> 00:11:07.320
So, a investigation is necessary

00:11:07.320 --> 00:11:13.000
to study in depth models to reduce biases.

00:11:13.000 --> 00:11:16.640
We cannot only use these language model
and directly

00:11:16.640 --> 00:11:19.520
use the outcome.

00:11:20.240 --> 00:11:22.520
Another problem with these model

00:11:23.520 --> 00:11:26.880
is that there aren’t too many dataset

00:11:27.360 --> 00:11:30.000
related to accessibility area.

00:11:30.000 --> 00:11:34.000
For instance, there a few label corpora

00:11:34.400 --> 00:11:36.760
to be used in training simplification,

00:11:37.200 --> 00:11:41.280
algorithms,
lexical or syntactic simplification

00:11:41.880 --> 00:11:43.600
in natural language processing.

00:11:43.600 --> 00:11:50.720
I work in cognitive accessibility
in a in Spanish to simplify text

00:11:50.720 --> 00:11:57.960
to plain language and easy
reading language. To carry out this case

00:11:57.960 --> 00:12:02.680
we had to create a corpus
with an expert in easy reading,

00:12:03.000 --> 00:12:05.680
with the participation of older people

00:12:06.360 --> 00:12:10.680
and with people with disability
intellectual disabilities

00:12:11.480 --> 00:12:14.480
because the current corpora
had been created

00:12:14.480 --> 00:12:19.520
with non expert in disability,
non expert in plain language

00:12:19.960 --> 00:12:24.000
and they haven't taken into account
the people with disability.

00:12:25.560 --> 00:12:30.360
Also an effort devoted to solving this

00:12:30.360 --> 00:12:34.040
scarcity of resources are required

00:12:34.040 --> 00:12:37.200
in language with low resources.

00:12:37.680 --> 00:12:42.600
English is the language with more developed
with many natural language processing.

00:12:42.600 --> 00:12:46.200
But others, such as Spanish, have hardly

00:12:46.200 --> 00:12:49.280
any resources. We need system

00:12:49.400 --> 00:12:54.720
trained for English language
to work for Spanish as well.

00:12:54.720 --> 00:12:58.680
And finally,
with the proliferation of GPT models

00:12:58.680 --> 00:13:03.160
with application
such as ChatGPT

00:13:03.480 --> 00:13:06.320
another problem to address

00:13:06.360 --> 00:13:09.760
is the regulation on ethical aspect of

00:13:10.800 --> 00:13:14.920
artificial intelligence.

00:13:15.840 --> 00:13:17.240
Okay, thank you so much.

00:13:17.240 --> 00:13:22.280
Lourdes, definitely
some very relevant challenges in there.

00:13:23.600 --> 00:13:26.160
Vikas, I’ll end this first round with you.

00:13:27.440 --> 00:13:30.360
Thank you Carlos. 
I’m Vikas Ashok

00:13:30.440 --> 00:13:34.600
from Old Dominion University
in Virginia United States.

00:13:35.160 --> 00:13:37.680
So I have been working researching

00:13:37.680 --> 00:13:41.480
in the area of accessible
computing for like ten years now.

00:13:42.000 --> 00:13:46.920
And my specialty focus area
is of people with visual disabilities.

00:13:47.320 --> 00:13:50.840
So I have mostly concentrated
on their accessibility

00:13:50.840 --> 00:13:54.000
as well as usability needs
when it comes to interacting

00:13:54.000 --> 00:13:56.280
with computer applications.

00:13:57.080 --> 00:14:01.560
So with the topic at hand,
which is accessible communication.

00:14:01.600 --> 00:14:04.560
So one of the projects
that I'm currently looking at

00:14:04.560 --> 00:14:10.080
is understandably of social media content
for people

00:14:10.600 --> 00:14:13.920
who listen to content such as,
you know, people who are blind.

00:14:14.640 --> 00:14:18.000
So listening,
you know, social media content,

00:14:18.000 --> 00:14:21.640
text is not the same as looking at it.

00:14:22.120 --> 00:14:25.320
So, even though the social media text

00:14:25.320 --> 00:14:31.160
is accessible, it's not necessarily
understandable because of presence

00:14:31.160 --> 00:14:34.880
of a lot of nonstandard language

00:14:34.880 --> 00:14:37.880
content in social media such as Twitter.

00:14:37.920 --> 00:14:39.600
Like people create their own words.

00:14:39.600 --> 00:14:44.040
They're very inventive there,
so they hardly follow any grammar.

00:14:44.880 --> 00:14:48.920
So text to speech systems
such as those used in screen

00:14:48.920 --> 00:14:52.920
readers cannot necessarily pronounce

00:14:52.960 --> 00:14:56.480
these out of
vocabulary words in the right way,

00:14:56.880 --> 00:14:59.920
because most of these words,
even though they're in text form,

00:15:00.360 --> 00:15:05.200
they're mostly intended for visual consumption, 
such as some kind of exaggeration

00:15:05.560 --> 00:15:10.680
where the letters are duplicated
just for some kind of additional effect.

00:15:11.000 --> 00:15:14.800
Sometimes even emotions are attached
to the text itself without any,

00:15:15.160 --> 00:15:17.480
you know, emoticons or anything else.

00:15:17.960 --> 00:15:22.640
So and sometimes they try
to phonetically match it,

00:15:22.640 --> 00:15:27.080
use a different spelling for a word
just for fun purposes.

00:15:27.240 --> 00:15:33.320
So this kind of fun as communication
has increased tremendously on social media

00:15:33.320 --> 00:15:39.000
and people are depending on social media
to understand or get news, even,

00:15:39.400 --> 00:15:42.760
you know, some kind of disaster news
or something happens anywhere.

00:15:42.760 --> 00:15:46.680
Some even, they first flock to the
social media to get it. So

00:15:47.680 --> 00:15:48.880
people who listen

00:15:48.880 --> 00:15:53.760
to content also should be able to easily
understand, so I’m focusing on that area

00:15:53.760 --> 00:15:57.480
how to use NLP to make this possible.

00:15:58.360 --> 00:16:02.360
Because even though this is not exactly
a question

00:16:02.440 --> 00:16:08.400
of accessibility in a conventional sense,
but it's more like accessibility

00:16:08.400 --> 00:16:12.960
in terms of being able to understand
the already accessible content.

00:16:12.960 --> 00:16:14.920
So that's one of the things.

00:16:14.920 --> 00:16:18.760
The other thing that we're looking at,
which is related to this panel

00:16:18.800 --> 00:16:22.600
is the... related to the bias, disability

00:16:22.600 --> 00:16:27.720
bias of natural language models,
especially those large language models.

00:16:28.480 --> 00:16:33.960
So unfortunately, these models
are reflective of the data it's trained on

00:16:34.600 --> 00:16:40.720
because most of the data associates words
that are used to describe people

00:16:40.720 --> 00:16:45.080
with disabilities somehow end up having
negative connotation.

00:16:45.360 --> 00:16:47.680
So they're using negative context.

00:16:48.040 --> 00:16:51.760
So it's nobody's telling these models
to learn it that way

00:16:52.320 --> 00:16:56.640
except that the documents
or the text corpus that these models

00:16:56.640 --> 00:17:01.240
are looking at inherently
put these words that are,

00:17:01.680 --> 00:17:06.200
you know, many times
not offensive into the negative category.

00:17:07.440 --> 00:17:10.080
So I'm looking at how we can counter

00:17:10.080 --> 00:17:14.160
this. One example is toxicity detection

00:17:14.320 --> 00:17:19.280
in discussion forums, online discussion
forums are very popular.

00:17:19.280 --> 00:17:22.280
People go there, sometimes anonymously,

00:17:22.280 --> 00:17:24.720
post content, interact with each other.

00:17:25.320 --> 00:17:28.440
And, you know,
some of the posts get flagged

00:17:28.560 --> 00:17:32.360
as, you know, toxic
or this get filtered out.

00:17:32.680 --> 00:17:35.800
So even if they're not toxic

00:17:36.160 --> 00:17:41.520
because of the use of certain words
to describe disabilities or something.

00:17:41.520 --> 00:17:43.760
So we want to avoid that.

00:17:43.760 --> 00:17:47.920
So how do we
how can we use NLP to not do that.

00:17:48.520 --> 00:17:52.920
So these two projects are pretty much
what's closely related to the panel.

00:17:53.400 --> 00:17:55.160
So stick to these.

00:17:55.160 --> 00:17:58.840
This session.

00:17:58.840 --> 00:17:59.240
Thank you, Vikas.

00:18:00.920 --> 00:18:05.720
I'll follow up with that
with what you've mentioned and Lourdes

00:18:05.800 --> 00:18:08.960
has also previously highlighted

00:18:08.960 --> 00:18:14.320
the disability bias
and... and I'm wondering

00:18:14.320 --> 00:18:18.240
if you have any ideas
and suggestions on how can

00:18:19.560 --> 00:18:22.320
NLP tools

00:18:22.920 --> 00:18:24.960
address such issues.

00:18:24.960 --> 00:18:27.960
I'm thinking,
for instance, text summarization tools,

00:18:27.960 --> 00:18:30.680
but also other also NLP tools.

00:18:30.680 --> 00:18:33.840
How can they help us
address issues of disability bias,

00:18:33.840 --> 00:18:37.920
but also how can they explore other aspects

00:18:37.920 --> 00:18:41.280
like accountability or personalization

00:18:41.280 --> 00:18:44.320
of in the case of

00:18:44.320 --> 00:18:45.480
text summaries?

00:18:45.480 --> 00:18:49.200
And how can I personalize a summary

00:18:49.200 --> 00:18:53.080
for specific audiences
for the needs of specific people?

00:18:54.200 --> 00:18:56.640
I'll start with you now Lourdes.

00:18:56.640 --> 00:18:57.600
OK.

00:18:58.680 --> 00:19:00.080
Text summarization is

00:19:00.080 --> 00:19:02.680
a natural language
processing task.

00:19:02.680 --> 00:19:04.920
Is a... is a great resource

00:19:06.720 --> 00:19:08.360
because improve cognitive

00:19:08.360 --> 00:19:14.080
accessibility in order to help people
people with disabilities to process alone

00:19:14.080 --> 00:19:16.880
and deduce text.

00:19:17.400 --> 00:19:20.280
Also, in the web
content accessibility guidelines

00:19:20.760 --> 00:19:25.200
following success criteria 3.1.5 
Reading Level

00:19:25.200 --> 00:19:29.720
the readable summary is a 

00:19:30.160 --> 00:19:33.000
resource that is recommended.

00:19:34.640 --> 00:19:37.920
But these task has challenges.

00:19:38.320 --> 00:19:42.680
Such us, bias, disability biasis.

00:19:43.040 --> 00:19:46.000
And the summaries generated

00:19:46.000 --> 00:19:49.320
are understandable 
for people with disability is

00:19:50.000 --> 00:19:54.520
is at a is understandable
for people with disability

00:19:55.080 --> 00:20:00.560
therefore some aspects must be taken
into account: is necessary

00:20:00.560 --> 00:20:06.320
to approach this task
which is summarize of the extractive type

00:20:07.520 --> 00:20:12.240
where the extracted sentences
can be modified with paraphrases

00:20:12.320 --> 00:20:17.400
resources and help understandability
and readability of the text.

00:20:18.520 --> 00:20:22.400
To summarize text
different input are required

00:20:22.960 --> 00:20:28.440
not only knowledge about the sequences
of words or other

00:20:28.800 --> 00:20:32.600
leads about sentences,

00:20:33.000 --> 00:20:37.320
but also about the target of audience
is important.

00:20:37.440 --> 00:20:40.480
Different type of user

00:20:40.480 --> 00:20:44.760
require different type
or personalization of summaries.

00:20:46.400 --> 00:20:50.520
It also a

00:20:50.760 --> 00:20:53.320
I think that this is

00:20:53.760 --> 00:20:56.280
it will be recommendable to include

00:20:56.480 --> 00:21:00.720
readability metric
in the summary generation process

00:21:01.240 --> 00:21:06.560
to ensure that the resulting
summary is minimally readable.

00:21:08.400 --> 00:21:09.000
For instance

00:21:09.000 --> 00:21:10.680
if

00:21:10.680 --> 00:21:15.880
we are in the context of a system
that provides summaries of public

00:21:15.960 --> 00:21:18.560
administration information 
for old people,

00:21:19.000 --> 00:21:22.880
it's necessary
to take into account that the summary

00:21:23.160 --> 00:21:26.600
must be in plain language,

00:21:26.760 --> 00:21:30.520
therefore in addition
to extract the relevant sentences

00:21:30.520 --> 00:21:35.760
and paraphrases it will be necessary
to include knowledge about of guideline

00:21:35.760 --> 00:21:39.720
of plain language to make
the text easier to read

00:21:40.560 --> 00:21:45.480
and finally corpora use

00:21:46.240 --> 00:21:50.280
to train natural language
processing system shall be tested

00:21:50.280 --> 00:21:54.880
with users it in order to attain
useful solution.

00:21:55.520 --> 00:21:59.520
Only then it will be possible
to obtain understandable summaries

00:21:59.520 --> 00:22:03.440
for the whole of society and the elderly

00:22:03.720 --> 00:22:07.920
and and with respect to accountability

00:22:09.120 --> 00:22:13.600
as as in every artificial intelligence
algorithm,

00:22:13.680 --> 00:22:16.200
these must be explainable

00:22:18.000 --> 00:22:21.120
so is necessary to respond to answer

00:22:21.680 --> 00:22:25.160
to questions such as how processing

00:22:25.160 --> 00:22:29.440
actually perform,
the limitation of the dataset

00:22:29.880 --> 00:22:34.960
used to train and test algorithms
and the outcomes of the model

00:22:36.040 --> 00:22:37.480
a therefore good

00:22:37.480 --> 00:22:41.640
data manager, management
and machine learning models

00:22:41.640 --> 00:22:42.640
trained in practice

00:22:42.640 --> 00:22:46.320
shall be promote 
to ensure quality results.

00:22:49.080 --> 00:22:51.000
And nothing else.

00:22:51.000 --> 00:22:52.880
Thank you Lourdes.

00:22:53.800 --> 00:22:58.920
Vikas, do you want to... even though we...
from what I understood,

00:22:58.920 --> 00:23:03.480
you don't work directly
with text summarization but still others

00:23:03.480 --> 00:23:07.720
aspects of disability
bias, accountability,

00:23:07.720 --> 00:23:11.040
personalization, impact what you're doing.

00:23:12.080 --> 00:23:16.760
Yeah, I mean I use a lot of text summarization
so I can add to it.

00:23:16.760 --> 00:23:21.520
So to add to what Lourdes said,

00:23:22.440 --> 00:23:26.560
simplification
is also as important as summarization

00:23:26.920 --> 00:23:32.240
because sometimes it's not just
summarizing or shortening the content

00:23:32.400 --> 00:23:35.840
to be consumed, but it's also making
it understandable, like I said.

00:23:36.280 --> 00:23:40.040
So that means that certain complex
sentence structures

00:23:40.040 --> 00:23:41.680
and some exotic words

00:23:41.680 --> 00:23:44.880
we need to replace them
with equal and easier

00:23:44.880 --> 00:23:48.080
to understand
more frequently used words.

00:23:48.240 --> 00:23:54.080
So there there is some work there that has
been gone into text simplification.

00:23:54.120 --> 00:23:57.840
We created some kind of summarization
in the special case.

00:23:58.160 --> 00:24:01.640
It's from the same language
to text from between the same language.

00:24:01.840 --> 00:24:05.800
So the input is a text
in the same language as the output text,

00:24:06.000 --> 00:24:09.720
except that the output text
is more readable, more understandable.

00:24:10.320 --> 00:24:12.360
So that is extremely important.

00:24:12.360 --> 00:24:13.440
The other thing is

00:24:14.400 --> 00:24:15.240
summarization.

00:24:15.240 --> 00:24:19.240
Most of them tend
to rely extractive summarization wherein

00:24:19.560 --> 00:24:24.080
they just pick certain sentences
from the original piece of text

00:24:24.760 --> 00:24:26.840
so that they don't have to worry about the

00:24:27.560 --> 00:24:30.800
grammatical correctness 
and proper sentence structures

00:24:31.320 --> 00:24:34.400
so that because they rely on humans

00:24:34.400 --> 00:24:37.320
who have written the text
in order to generate the summaries.

00:24:37.760 --> 00:24:40.080
So I can speak about

00:24:40.440 --> 00:24:44.840
how summarization needs to be personalized
in certain way for certain groups,

00:24:44.960 --> 00:24:47.560
especially for people
with visual disabilities.

00:24:47.920 --> 00:24:51.720
So what I have noticed in
some of my studies is that

00:24:52.640 --> 00:24:56.480
even though they can hear it,
they don't necessarily understand it

00:24:56.480 --> 00:24:59.840
because the writing
is sort of visual in the sense

00:24:59.840 --> 00:25:03.040
it needs to need you
to be visually imaginative.

00:25:03.600 --> 00:25:06.560
So what is the alt... 
the non-visual alternative

00:25:07.240 --> 00:25:09.480
for such kind of text?

00:25:09.840 --> 00:25:15.080
So how do you summarize the text that
includes a lot of visual elements to it.

00:25:15.280 --> 00:25:18.080
So how do you convert it into non

00:25:18.840 --> 00:25:21.520
equal non-visual explanations?

00:25:21.640 --> 00:25:26.000
This necessarily goes
beyond the extractive summarization.

00:25:26.000 --> 00:25:27.760
You cannot just pick and choose,

00:25:27.760 --> 00:25:31.760
so you need to replace the wordings
in the sentence.

00:25:31.760 --> 00:25:36.120
By other wordings that they can understand
and some of the text,

00:25:36.120 --> 00:25:37.960
you know, these days, especially

00:25:37.960 --> 00:25:42.240
the articles, news articles and all,
they don't come purely as text.

00:25:42.240 --> 00:25:44.920
They're sort of multi-modal in the sense

00:25:45.240 --> 00:25:48.000
there are pictures, there are GIFs everything.

00:25:48.000 --> 00:25:51.200
And the text sort of refers
to these pictures

00:25:52.120 --> 00:25:56.840
so that this is another problem
because then it becomes highly visual.

00:25:56.840 --> 00:26:00.240
So you have to take
some of the visual elements

00:26:00.240 --> 00:26:03.920
of the picture, probably through computer
vision techniques or something,

00:26:03.920 --> 00:26:08.760
and then inject it into the text
in order to make it more self-sufficient

00:26:08.760 --> 00:26:12.800
and understandable
for people who cannot see the images.

00:26:13.760 --> 00:26:17.240
So that's my take on it.

00:26:17.280 --> 00:26:23.240
Yeah, that's a very good point
about the multimedia information

00:26:23.240 --> 00:26:26.320
and how do we summarize everything
into text.

00:26:26.320 --> 00:26:28.920
Yeah, that's a great point.

00:26:28.920 --> 00:26:31.560
Chaohai, your take on this?

00:26:31.560 --> 00:26:33.240
Oh yes. Yeah.

00:26:33.240 --> 00:26:36.880
But we don't know much experience
in text summarization.

00:26:36.880 --> 00:26:42.640
Most our research is focused
on the residual AAC and interlinking

00:26:42.680 --> 00:26:46.600
and the AAC generation,
but we do how well the project

00:26:47.040 --> 00:26:49.200
involved part of the text summarization.

00:26:50.040 --> 00:26:53.240
We construct a knowledge graph

00:26:53.640 --> 00:26:56.520
for e-learning platform
and that we need to

00:26:57.680 --> 00:26:58.880
extract

00:26:59.280 --> 00:27:04.520
the text summarization from lecture notes
to, make it easier

00:27:04.520 --> 00:27:07.680
and accessible for people, students

00:27:07.680 --> 00:27:11.160
with disabilities. So,

00:27:11.160 --> 00:27:15.440
so based on that project,

00:27:15.440 --> 00:27:19.440
what we learned is that text
summarization is very difficult task

00:27:20.880 --> 00:27:24.480
in NLP because these are
highly dependent on the text

00:27:25.600 --> 00:27:27.240
context domain or

00:27:27.240 --> 00:27:31.040
target audience and even the goal summary.

00:27:31.040 --> 00:27:36.680
For example, in our scenario,
we want to have the summary of

00:27:37.960 --> 00:27:39.880
each lecture notes,

00:27:39.880 --> 00:27:43.480
but we a very long transcripts
in that lecture.

00:27:43.920 --> 00:27:46.280
So we use a few

00:27:47.480 --> 00:27:49.680
text summarization models to generate

00:27:52.360 --> 00:27:55.200
the summaries, but the outcome is not good

00:27:56.360 --> 00:27:56.920
somewhere.

00:27:56.920 --> 00:27:59.120
It is mainly as

00:28:00.280 --> 00:28:03.520
Vikas just said, some of the text

00:28:03.520 --> 00:28:07.880
summarization is just pick
some of the text and replace

00:28:07.880 --> 00:28:12.960
some of the words that say so
or even some that doesn't make sense.

00:28:13.360 --> 00:28:16.520
So that's the one.

00:28:16.520 --> 00:28:20.640
One problem
we identified in text summarization

00:28:21.120 --> 00:28:24.000
and we also have some

00:28:26.000 --> 00:28:28.200
method to

00:28:28.920 --> 00:28:29.800
to read

00:28:31.440 --> 00:28:33.360
because we we

00:28:33.360 --> 00:28:36.400
we need to personalize
because the project is

00:28:37.240 --> 00:28:41.280
related to the adaptive learning
for individual students.

00:28:41.280 --> 00:28:44.080
We need a personalization
for each student. So

00:28:45.200 --> 00:28:47.920
personalization could be

00:28:47.920 --> 00:28:51.760
customized, adapted to user need.

00:28:52.240 --> 00:28:55.760
But this is actually

00:28:55.760 --> 00:28:58.920
can can be improved the ways

00:29:01.000 --> 00:29:04.000
users’ personal preferences

00:29:04.400 --> 00:29:06.720
or feedback and

00:29:08.200 --> 00:29:11.040
and also allow user to set the

00:29:12.600 --> 00:29:14.800
summary goal and...

00:29:14.800 --> 00:29:20.120
and also the simplification
is very important because some students

00:29:20.360 --> 00:29:25.320
may have cognitive disability
or or other type of disability.

00:29:25.320 --> 00:29:28.200
They need to have simplified or blend

00:29:28.760 --> 00:29:33.720
plain language. Yet.

00:29:33.720 --> 00:29:35.760
Yeah, I think that's mainly what we

00:29:37.200 --> 00:29:40.840
have for text summarization.

00:29:40.840 --> 00:29:41.520
Thank you, Chaohai.

00:29:41.600 --> 00:29:43.920
Thank you. Uh.

00:29:43.920 --> 00:29:48.480
Okay, so let's move on to what we started
with the challenges

00:29:48.480 --> 00:29:53.240
and now I would like to move on
to the future perspectives.

00:29:53.240 --> 00:29:58.440
How do... what are the breakthroughs
that you see happening

00:29:58.960 --> 00:30:01.280
promoted by the use of NLP

00:30:02.080 --> 00:30:04.320
for accessible communication.

00:30:04.680 --> 00:30:06.920
And I'll we'll start with you now, Vikas.

00:30:10.960 --> 00:30:12.840
So my

00:30:12.840 --> 00:30:16.200
perspective
is that there's plenty of NLP,

00:30:16.200 --> 00:30:20.120
you know tools out there already
that haven’t been exploited

00:30:20.120 --> 00:30:24.760
to the fullest extent to address
accessibility and usability issues.

00:30:25.280 --> 00:30:27.960
The growth in NLP techniques

00:30:27.960 --> 00:30:33.080
and methods that has been extremely steep
in the recent years

00:30:33.080 --> 00:30:37.080
and the rest of us in different
fields are trying to catch up.

00:30:37.560 --> 00:30:39.840
I mean, still,
there is a lot to be explored

00:30:40.400 --> 00:30:43.320
as to how they can be used to address

00:30:43.320 --> 00:30:45.880
real world accessibility problems,

00:30:46.360 --> 00:30:51.120
and we are in the process of doing that,
I would say so

00:30:51.120 --> 00:30:53.320
text summarization is one thing

00:30:54.240 --> 00:30:57.000
that we discussed already,
which can be exploited

00:30:57.000 --> 00:31:01.840
in a lot of scenarios
to improve the efficiency

00:31:01.840 --> 00:31:05.880
of computer interaction
for people with disabilities.

00:31:06.240 --> 00:31:09.840
But the main problem, as we discussed
not only in this panel

00:31:09.840 --> 00:31:12.120
but also on other panels, is the data.

00:31:12.120 --> 00:31:16.560
So for some languages
there is enough ... corpus

00:31:16.560 --> 00:31:19.960
where the translation is good,
because the translation

00:31:19.960 --> 00:31:23.280
essentially depends on
how much data you have trained on.

00:31:23.640 --> 00:31:27.200
But for some pair of
languages it may not be that

00:31:28.200 --> 00:31:28.800
easy.

00:31:28.800 --> 00:31:31.680
Or even if it
does something may not be that accurate.

00:31:31.680 --> 00:31:33.800
So that may be a problem.

00:31:33.800 --> 00:31:36.920
And then the biggest area where I see,

00:31:37.480 --> 00:31:41.160
which can be very useful for solving

00:31:41.160 --> 00:31:46.560
many accessibility problems
is the improvement in dialog systems.

00:31:46.560 --> 00:31:49.560
So natural language dialog is more like

00:31:49.560 --> 00:31:52.000
a really intuitive interface for many

00:31:53.360 --> 00:31:56.840
users, including many people
with disabilities.

00:31:57.360 --> 00:32:00.240
So those who have physical impairments
which

00:32:01.280 --> 00:32:05.040
prevent them from conveniently
using the keyboard or the mouse

00:32:05.040 --> 00:32:08.080
and those are blind
who have to use screen readers,

00:32:08.080 --> 00:32:11.040
which is time
consuming, is known to be time consuming.

00:32:11.040 --> 00:32:15.080
So dialog assistants are, I would say

00:32:16.320 --> 00:32:17.720
under-explored...
they're still exploring it.

00:32:17.720 --> 00:32:21.840
We see that commercialization is going on
like smartphones and all,

00:32:22.080 --> 00:32:26.240
but still it's at the level
of some high level interaction

00:32:26.240 --> 00:32:30.840
like setting alarms or turning on lights
and answering some questions.

00:32:31.280 --> 00:32:34.760
But what about using that to interact
with applications

00:32:34.760 --> 00:32:36.760
in the context of an application?

00:32:37.360 --> 00:32:39.480
So if I see a play,

00:32:41.040 --> 00:32:41.760
I had a user

00:32:41.760 --> 00:32:44.760
comment to this particular document.

00:32:44.760 --> 00:32:49.280
It's in word or Google Docs. So can an
assistant spoken

00:32:49.280 --> 00:32:52.200
dialog assistant
understand that an automated

00:32:53.360 --> 00:32:55.440
means so this kind of automation

00:32:56.240 --> 00:33:01.240
will sort of address,
I feel will address many of the issues

00:33:01.240 --> 00:33:04.640
that people face
interacting with digital content.

00:33:04.640 --> 00:33:09.080
So that's one of the things I would say
we can use NLP for.

00:33:09.480 --> 00:33:15.120
The other thing is the increased
availability of large language

00:33:15.120 --> 00:33:20.280
models, Pre-trained models
like one Lourdes mentioned, like GPT,

00:33:20.840 --> 00:33:25.680
which is essentially transformer decoder
or generator based model.

00:33:25.680 --> 00:33:28.040
Then there's also Bert,
which encoder based.

00:33:28.440 --> 00:33:32.040
So these help us, you know,

00:33:32.120 --> 00:33:36.920
in a way that we don't need large
amounts of data to solve problems

00:33:36.920 --> 00:33:40.320
because they're already pre-trained
on a large amount of data.

00:33:40.800 --> 00:33:44.560
So what we would need are kind of small

00:33:44.760 --> 00:33:49.800
data sets that are more fine tuned
towards the problem we are addressing.

00:33:50.520 --> 00:33:53.960
So the datasets,
they're accessibility datasets.

00:33:53.960 --> 00:33:57.600
They're I think there needs to be
a little bit more investment

00:33:58.480 --> 00:34:02.960
doesn't have to be that big
because the large language models

00:34:02.960 --> 00:34:06.640
already take care of most of the language
complexity.

00:34:06.760 --> 00:34:08.200
It's more like fine tuning

00:34:09.160 --> 00:34:10.560
the problem at hand.

00:34:10.560 --> 00:34:14.400
So that's where I think
some effort should go.

00:34:14.400 --> 00:34:18.800
And once we do that, obviously
we can fine tune and solve the problems

00:34:18.800 --> 00:34:23.640
and then there is a tremendous enhancement
or advancement

00:34:23.640 --> 00:34:27.920
in transport learning techniques
which we can exploit that as well,

00:34:28.440 --> 00:34:32.280
in order to not do stuff from scratch,
instead borrow

00:34:32.280 --> 00:34:35.640
some things that are already there
for something different.

00:34:36.320 --> 00:34:38.280
I mean, similar problem.

00:34:38.280 --> 00:34:43.880
So, so there is a lot to be explored,
but we haven't done that yet.

00:34:43.880 --> 00:34:45.920
So there's plenty of opportunity

00:34:45.920 --> 00:34:48.960
for research
using NLP expertise for

00:34:49.560 --> 00:34:53.400
problems in accessible communication,
especially.

00:34:53.400 --> 00:34:56.160
Yes, definitely
some exciting avenues there.

00:34:57.120 --> 00:35:00.680
So, Chaohai, can we have your take on this?

00:35:01.120 --> 00:35:04.080
What are your breakthroughs?

00:35:04.200 --> 00:35:04.960
OK.

00:35:04.960 --> 00:35:08.560
Just listened them,
Vikas, I totally agree with him.

00:35:08.880 --> 00:35:09.240
He's

00:35:11.160 --> 00:35:13.200
all opinions and

00:35:13.200 --> 00:35:15.320
for... for my research

00:35:15.720 --> 00:35:18.480
because I've mainly worked on AAC, 
so, currently,

00:35:19.160 --> 00:35:22.160
so I would take AAC for example.

00:35:22.920 --> 00:35:26.120
So the future perspective for AAC,

00:35:27.040 --> 00:35:29.200
NLP for AAC, I think the first of

00:35:29.240 --> 00:35:32.920
that will be the personalized adaptive

00:35:34.160 --> 00:35:37.480
communication for each individual,
because

00:35:39.240 --> 00:35:42.520
each individual
has their own communication,

00:35:43.640 --> 00:35:45.760
their own way to communicate
with each other

00:35:46.120 --> 00:35:49.200
and NLP techniques can be used to make

00:35:50.320 --> 00:35:52.720
this communication more accessible,

00:35:52.920 --> 00:35:56.480
more personalized
and adaptive based on their

00:35:58.000 --> 00:36:01.200
personal preferences of feedback.

00:36:02.440 --> 00:36:06.520
So this can can be used to

00:36:06.840 --> 00:36:10.320
for personalize the AAC symbols so

00:36:11.320 --> 00:36:14.440
currently, AAC users

00:36:14.440 --> 00:36:19.080
they just using some standard
AAC symbol set for

00:36:19.160 --> 00:36:24.480
their daily communications,
so how can we use NLP to

00:36:25.680 --> 00:36:29.520
and generic and the generic AI models

00:36:30.240 --> 00:36:34.720
to create a more customized
personalized AAC symbols

00:36:36.080 --> 00:36:37.440
that's which

00:36:37.440 --> 00:36:40.800
which you could be have ability
to adapt to the

00:36:42.720 --> 00:36:45.240
individual's
unique cultural and social needs.

00:36:45.560 --> 00:36:48.240
I think that's one potentially

00:36:49.280 --> 00:36:52.280
contribute to the AAC users.

00:36:52.800 --> 00:36:56.160
The second one will be accessible multi

00:36:56.320 --> 00:36:59.440
modal communication

00:37:00.120 --> 00:37:01.720
because a

00:37:02.000 --> 00:37:03.920
that NLP techniques

00:37:03.920 --> 00:37:07.560
they have the potential to enhance this

00:37:08.520 --> 00:37:11.600
accessible communication by improving

00:37:12.960 --> 00:37:16.320
interoperability in training data

00:37:17.000 --> 00:37:22.160
and the between the verbal language
sign language and that AAC so data

00:37:22.160 --> 00:37:26.280
interoperability could provide
more high quality training data for this

00:37:28.120 --> 00:37:28.960
language with

00:37:28.960 --> 00:37:31.040
elastic set and

00:37:32.200 --> 00:37:35.400
additionally,

00:37:35.400 --> 00:37:39.720
it can provide the ability to translate

00:37:40.280 --> 00:37:43.200
different communication models

00:37:43.800 --> 00:37:46.280
and to make it more accessible
and inclusive.

00:37:47.480 --> 00:37:50.280
So in AAC, so we can have

00:37:51.440 --> 00:37:55.080
multiple AAC symbol sets can be link

00:37:55.640 --> 00:37:58.200
mapped and interlinked by NLP models

00:37:58.680 --> 00:38:03.200
and this can be contribute
to the translation between the AAC to AAC

00:38:03.520 --> 00:38:07.800
and AAC to text, AAC to some language
and vice versa.

00:38:08.480 --> 00:38:13.880
Yeah, that's the,
the the second aspect I think about.

00:38:13.880 --> 00:38:15.240
And then the third one is the

00:38:17.240 --> 00:38:19.720
AI assistant
communication that Vikas

00:38:19.720 --> 00:38:23.760
just, just talk about the ChatGPT.

00:38:23.760 --> 00:38:26.760
So with this,

00:38:26.760 --> 00:38:30.080
this large language model has been trained

00:38:30.080 --> 00:38:32.840
by this big companies and

00:38:33.960 --> 00:38:35.680
and they have been widely 
spreading on social media.

00:38:35.680 --> 00:38:38.560
So how

00:38:38.560 --> 00:38:40.680
how to using this

00:38:40.680 --> 00:38:42.920
this trained large

00:38:43.400 --> 00:38:48.360
language models incorporated
with other applications and then can use it

00:38:48.360 --> 00:38:53.560
for a more accessible communication
to help people with disabilities.

00:38:54.000 --> 00:38:56.800
That's that's another

00:38:57.000 --> 00:38:57.560
future

00:38:57.560 --> 00:38:59.200
we are looking for.

00:38:59.200 --> 00:39:01.320
The last one 
that I'm going to talk about

00:39:01.560 --> 00:39:04.880
is more regarding the AAC
because it's quite expensive.

00:39:05.160 --> 00:39:07.960
So affordability is very important

00:39:08.640 --> 00:39:13.080
and it can be achieved by the NLP

00:39:13.320 --> 00:39:18.120
or AI. That's what I mentioned
that we are currently looking into

00:39:18.640 --> 00:39:23.320
how to turn images into symbols
and how to generate

00:39:24.640 --> 00:39:28.200
AAC symbols automatically by using

00:39:29.520 --> 00:39:33.720
image generative AI models
like stable diffusion. So

00:39:34.840 --> 00:39:37.960
so that's
the another future we are looking forward

00:39:37.960 --> 00:39:41.280
how to reduce the cost
for accessible communication.

00:39:42.000 --> 00:39:44.280
Yeah. Thank you.

00:39:44.280 --> 00:39:44.640
Thank you,

00:39:44.640 --> 00:39:47.520
Chaohai. Definitely a relevant point.

00:39:47.520 --> 00:39:52.040
Reducing costs of getting data
and all of that.

00:39:52.320 --> 00:39:54.120
That's important everywhere.

00:39:54.120 --> 00:39:57.120
So, Lourdes, what are you looking for

00:39:57.360 --> 00:40:01.840
in the near future?

00:40:01.840 --> 00:40:05.880
And you are muted.

00:40:05.880 --> 00:40:11.920
So as we have mentioned before,
there are two trends that are the

00:40:12.360 --> 00:40:16.680
the appearance of new and better
language model than the previous one

00:40:17.120 --> 00:40:19.680
working in these these new models

00:40:20.120 --> 00:40:25.080
and to reduce the disability biases.

00:40:25.080 --> 00:40:29.040
Also I am going to list a specific
natural language processing

00:40:29.040 --> 00:40:31.800
task and data application

00:40:32.080 --> 00:40:34.320
that I will work in the coming year.

00:40:35.480 --> 00:40:39.960
And one of them is accessibility
to domain specific.

00:40:39.960 --> 00:40:41.680
task, such as, health

00:40:42.760 --> 00:40:44.720
the the

00:40:44.720 --> 00:40:47.960
health language is highly
demanded need

00:40:48.960 --> 00:40:51.800
but patients have problems understanding

00:40:51.800 --> 00:40:57.240
information about their health
condition, diagnosis, treatment

00:40:57.520 --> 00:41:02.560
and natural language processing
method could improve their understanding

00:41:02.560 --> 00:41:05.440
of health related documents.

00:41:06.400 --> 00:41:10.800
Similar sample appear in
legal and financial documents,

00:41:10.800 --> 00:41:14.880
the language of administration,
e-government...

00:41:15.480 --> 00:41:18.960
Current natural language
processing technology

00:41:18.960 --> 00:41:27.240
that simplifies and summarizes
this could help in the roadmap.

00:41:27.240 --> 00:41:29.480
And another

00:41:31.000 --> 00:41:33.720
line is speech to text

00:41:35.040 --> 00:41:37.800
speech to text will be a relevant area

00:41:37.800 --> 00:41:40.800
of research in the field
of virtual meetings

00:41:41.720 --> 00:41:46.440
in order to facilitate accessible,
accessible communication by generation

00:41:46.440 --> 00:41:50.120
of summaries of meeting
as well as minutes

00:41:51.280 --> 00:41:53.640
in plain language.

00:41:53.960 --> 00:41:56.320
Another topic is the

00:41:57.240 --> 00:42:00.600
integration of natural language
processing

00:42:00.600 --> 00:42:03.960
method into the design

00:42:03.960 --> 00:42:07.320
and development of
multimedia user interface

00:42:08.400 --> 00:42:12.120
is necessary to face accessible
accessible communication

00:42:12.120 --> 00:42:17.000
from a multi-disciplinary approach
between different areas

00:42:17.000 --> 00:42:20.520
such as, human computer interaction
software engineering

00:42:20.840 --> 00:42:24.360
and natural language processing.

00:42:24.360 --> 00:42:29.280
Finally, another
issue is advancing

00:42:29.280 --> 00:42:33.400
application in smart assistant
in natural language processing

00:42:33.920 --> 00:42:37.440
method to support people with disabilities
and the elderly.

00:42:38.520 --> 00:42:41.480
Assist them in their daily tasks,

00:42:41.760 --> 00:42:46.560
and promote active living.

00:42:46.560 --> 00:42:47.920
Okay thank you so much, Lourdes,

00:42:47.920 --> 00:42:52.120
and everyone of you
for for those perspectives.

00:42:52.520 --> 00:42:55.480
I guess we still have 5 minutes

00:42:55.480 --> 00:43:00.520
more in this session,
so I will risk another question

00:43:00.520 --> 00:43:05.040
and I will ask you to
to try to to be brief on this one.

00:43:05.040 --> 00:43:09.720
But the the need for data

00:43:09.720 --> 00:43:13.720
was common across all your interventions.

00:43:13.720 --> 00:43:17.480
And if we go back to the previous
panel, also,

00:43:17.480 --> 00:43:21.360
it was brought up by every every panelist.

00:43:21.360 --> 00:43:23.760
So yeah, we need data.

00:43:25.000 --> 00:43:26.880
What are your thoughts and

00:43:26.880 --> 00:43:29.600
how can we make it easier

00:43:30.120 --> 00:43:32.600
and to collect more data

00:43:32.880 --> 00:43:37.120
for the specific aspect
of accessible communication?

00:43:37.120 --> 00:43:38.960
Because, we communicate a lot, right?

00:43:38.960 --> 00:43:41.280
Technology has

00:43:41.680 --> 00:43:43.720
allowed us and open up

00:43:44.320 --> 00:43:46.760
several channels
to where we can communicate

00:43:46.960 --> 00:43:49.960
even when we're not co-located.

00:43:50.520 --> 00:43:52.600
So yeah, every one of us is

00:43:53.160 --> 00:43:56.200
different points of the planet
and we are communicating right now.

00:43:56.360 --> 00:44:01.440
Technology improved those that
that possibility a lot.

00:44:01.880 --> 00:44:04.680
And however

00:44:04.680 --> 00:44:07.320
we always hear this, 
we need more data.

00:44:07.320 --> 00:44:08.520
We can't get data.

00:44:08.520 --> 00:44:11.760
So how do you think we can get more data

00:44:13.160 --> 00:44:15.240
is and of course we need

00:44:15.240 --> 00:44:17.760
the data to train these models, but

00:44:18.280 --> 00:44:21.760
can't we also rely on these models
to generate data?

00:44:22.960 --> 00:44:25.320
So let me just

00:44:25.320 --> 00:44:28.560
drop this on you now and

00:44:30.120 --> 00:44:31.320
any of you want

00:44:31.320 --> 00:44:37.120
to go first?

00:44:37.120 --> 00:44:40.160
I can go first. Okay. Yeah. Yeah.

00:44:40.160 --> 00:44:43.320
Because we have actually working on open

00:44:43.320 --> 00:44:46.240
data four years ago before, I mean

00:44:47.440 --> 00:44:50.680
the AI and the data science
because when I started my PhD

00:44:50.840 --> 00:44:55.720
we working on the open data and
there is a open data initiative in UK.

00:44:56.040 --> 00:44:59.720
So we want to open our data
and government data

00:45:01.120 --> 00:45:05.120
and the, and the public transport data
and that's

00:45:05.120 --> 00:45:09.200
how long working on public transportation
with accessibility needs.

00:45:09.200 --> 00:45:11.400
So there's a lack of data.

00:45:11.520 --> 00:45:13.240
At the beginning of my Ph.D.

00:45:13.240 --> 00:45:17.320
so few years later, the still
lack of the accessibility information data

00:45:18.000 --> 00:45:19.240
on this data.

00:45:19.240 --> 00:45:23.400
So the I think the so how can we
how is this

00:45:23.760 --> 00:45:28.920
I mean, the accessibility area,
how can we have such a data

00:45:29.320 --> 00:45:32.000
to to have to train our model?

00:45:32.000 --> 00:45:36.320
I mean, the first advice,
what I used to do

00:45:36.320 --> 00:45:40.560
with public transport data is I mapped all

00:45:41.520 --> 00:45:42.760
available data

00:45:42.760 --> 00:45:46.400
into a larger dataset. That's incurred

00:45:46.480 --> 00:45:50.120
a lot of labor work
like the cleaning data integration

00:45:50.120 --> 00:45:54.560
and all this
method to make data available.

00:45:55.120 --> 00:45:57.560
That's the first first approach.

00:45:57.840 --> 00:46:00.960
The second is we think about how can we

00:46:02.200 --> 00:46:05.360
contribute like

00:46:05.360 --> 00:46:08.160
a data repository or something

00:46:08.160 --> 00:46:13.120
like an image net or word net
that we can collaborate to together

00:46:13.160 --> 00:46:15.520
to contribute the identified

00:46:16.520 --> 00:46:19.320
data related to accessibility research.

00:46:20.160 --> 00:46:25.680
I think that that's we can as a community,
we can create such a universal

00:46:26.000 --> 00:46:28.320
repository or, or

00:46:30.160 --> 00:46:32.200
yeah, some kind of data initiative

00:46:32.200 --> 00:46:35.480
that we can working on
accessibility research.

00:46:36.240 --> 00:46:38.920
And the third approach 
is that definitely 

00:46:39.000 --> 00:46:42.440
that we can generate
the data based on the small data.

00:46:42.440 --> 00:46:45.400
We can be using

00:46:45.840 --> 00:46:48.720
generative AI model to generate more,

00:46:49.840 --> 00:46:53.000
but to
do the question is, is that data reliable?

00:46:53.080 --> 00:46:58.240
The data to generate generate enough,
or is that then the bias?

00:46:58.640 --> 00:47:01.600
So yeah, that's my my conclusion.

00:47:01.960 --> 00:47:02.680
Thank you.

00:47:03.400 --> 00:47:04.440
Yes, exactly.

00:47:04.440 --> 00:47:07.440
That's the... the big question mark.

00:47:07.440 --> 00:47:08.200
Right.

00:47:08.360 --> 00:47:11.520
Is that synthetic data
reliable or not, so

00:47:12.000 --> 00:47:14.440
Vikas or Lourdes
do you want to add something?

00:47:15.480 --> 00:47:18.000
So yeah I mean I have used synthetic data

00:47:18.120 --> 00:47:21.000
before based

00:47:21.040 --> 00:47:24.240
the little bit of real data
and in some cases

00:47:24.240 --> 00:47:26.440
you can generate synthetic data.

00:47:26.680 --> 00:47:29.640
So one of the things I had to do
was extract

00:47:30.400 --> 00:47:32.440
user comments in documents.

00:47:33.280 --> 00:47:37.480
Most of these word processing applications
allow you to post comments

00:47:37.960 --> 00:47:42.800
to the right for your collaborators to look at
and then, you know, address them.

00:47:43.240 --> 00:47:47.360
So automatically extracting
that I had to generate synthetic data

00:47:47.360 --> 00:47:51.800
because obviously you have few documents
with collaborative comments.

00:47:52.320 --> 00:47:54.800
So the appearance there is like, okay

00:47:54.840 --> 00:47:58.360
comments will appear somewhere
on the right side, right corner,

00:47:59.080 --> 00:48:02.400
which will have some text in it
with a few sentences.

00:48:02.680 --> 00:48:04.680
So there are some characteristics.

00:48:04.680 --> 00:48:07.640
So in those cases
we generated synthetic data.

00:48:07.640 --> 00:48:09.960
We train the machine learning model.

00:48:09.960 --> 00:48:13.720
It was pretty accurate on this data,
which was like real data.

00:48:14.520 --> 00:48:16.200
So... exploit...

00:48:16.200 --> 00:48:19.320
In some cases you can exploit the

00:48:19.680 --> 00:48:23.840
way data will appear
and then generate the synthetic data.

00:48:23.920 --> 00:48:27.240
But in many cases it may not be possible.

00:48:27.480 --> 00:48:30.480
Like for the project
I mentioned, social media,

00:48:30.480 --> 00:48:34.000
where text contains
a lot of nonstandard words.

00:48:34.600 --> 00:48:39.800
Simply replacing the nonstandard words
with synonyms may not do the job

00:48:39.880 --> 00:48:43.920
because then you take the fun aspect
away from social media.

00:48:44.280 --> 00:48:47.320
Like, it should be as fun and entertaining

00:48:47.320 --> 00:48:51.120
when you listen to social media text 
as it is when you look at it.

00:48:51.640 --> 00:48:54.360
So. So you have to do some kind of clever,

00:48:55.360 --> 00:48:56.880
you know, replacement.

00:48:56.880 --> 00:49:01.040
And for that you need some kind of expert
human expert going there and

00:49:03.000 --> 00:49:04.200
doing that.

00:49:04.200 --> 00:49:07.920
So crowdsourcing
I think is one way to get data

00:49:07.920 --> 00:49:11.040
quickly and it's pretty reliable.

00:49:11.440 --> 00:49:14.200
And see, I've seen in the NLP community

00:49:14.240 --> 00:49:19.560
like NLP papers that appear in ACL
and they rely heavily on the Amazon

00:49:20.160 --> 00:49:23.280
Mechanical Turk and other online

00:49:26.160 --> 00:49:29.120
incentivized data collection mechanisms.

00:49:29.560 --> 00:49:31.720
So that I think is one thing.

00:49:31.720 --> 00:49:35.640
The other thing I do
know, you know, in my classes

00:49:35.640 --> 00:49:39.840
especially, I get the students
to help each other out, collect the data

00:49:40.160 --> 00:49:43.560
so it doesn't have to be that intensive
every day.

00:49:43.560 --> 00:49:47.400
If they just even one student
collects like ten data points

00:49:47.880 --> 00:49:52.560
over the semester, it would be like
enough data for a lot of things.

00:49:52.560 --> 00:49:56.960
So you know, in each other projects
and in the end of the course

00:49:57.120 --> 00:50:00.320
pretty much they'll have a
lot of data for research. So

00:50:01.320 --> 00:50:02.120
you know,

00:50:02.120 --> 00:50:06.040
everybody can contribute
in a way and students

00:50:06.040 --> 00:50:10.200
especially are much more reliable
because they are familiar

00:50:10.200 --> 00:50:14.840
with the mechanisms
how to label collect data.

00:50:14.840 --> 00:50:18.520
And also they can understand
how things work as well.

00:50:18.520 --> 00:50:22.760
So, it’s like a win-win.

00:50:22.760 --> 00:50:25.000
Okay, yeah, thanks for that contribution.

00:50:25.480 --> 00:50:26.880
Good suggestion.

00:50:26.880 --> 00:50:31.040
And, Lourdes, we are really running out of time

00:50:31.040 --> 00:50:34.120
but if you still want to intervene,

00:50:34.120 --> 00:50:36.200
I can give you a couple of minutes.

00:50:37.240 --> 00:50:37.760
Okay.

00:50:37.760 --> 00:50:40.720
Only a I think that also

00:50:40.720 --> 00:50:46.080
we don't find many
we need a few data, but in my vision

00:50:46.080 --> 00:50:49.240
is also negative because obtaining

00:50:49.240 --> 00:50:52.640
the dataset is expensive.

00:50:52.680 --> 00:50:56.760
An in accessible communication,
I work in simplification.

00:50:57.960 --> 00:50:59.880
these data must be prepared

00:50:59.880 --> 00:51:01.960
by the expert in accessibility

00:51:03.360 --> 00:51:05.720
is important as these data

00:51:07.240 --> 00:51:11.400
is validated
by people with disability

00:51:11.680 --> 00:51:16.160
and use
plain language resources

00:51:16.880 --> 00:51:20.040
and then it is a problem

00:51:20.520 --> 00:51:25.640
to obtain data with quality.

00:51:26.880 --> 00:51:28.840
Okay, thank you so much,

00:51:28.840 --> 00:51:29.320
Lourdes.

00:51:29.320 --> 00:51:34.800
And thanks a very big
thank you to the three of you, Chaohai,

00:51:34.800 --> 00:51:36.680
Vikas and Lourdes. It

00:51:36.760 --> 00:51:38.760
was a really interesting panel

00:51:39.480 --> 00:51:42.040
thank you so much for
for your availability and