See also: IRC log
<Fran> Hi
(i prefer to skip the introduction, sitting in shared office at the moment)
I think 412.608 needs to be muted
Scott, your voice is still chopped up.
<epichler> lost audio
scott, try again
much better now
oh wait, there is the choppiness again
<Stephen> still very choppy
<epichler> am getting very intermittent audio
<rboyce> lost all sound
<scribe> scribenick: matthias_samwald
<mscottm> I will work on my audio while Francois goes ahead
<mscottm> could you hear me ask Francois to go ahead?
<rboyce> no, calling in again now
francois: i am here to introduce
the instance matching track at OAEI
... "ontology alignment evaluation initiative"
... has been going on for some years. running as workshop at
ISWC.
... we have been proposing a track about matching instance data
coming from "real" datasets, linked data.
... this year we want to focus on two points: 1) scale, 2)
different datasets, those without explicit ontology
<mscottm> Sorry about the sound problems - I was in a car accident last night and apparently my computer didn't survive completely intact.. :( I myself received an airbag smack to the face and chest.
francois: we want to focus on automated interlinking.
(sounds bad, scott!)
scribe: we want to use manually curated links as a reference point to judge the quality of automated interlinking tools.
<Fran> http://oaei.ontologymatching.org/
<Fran> http://oaei.ontologymatching.org/2009/
scott: have you chosen a language to represent mappings?
francois: OAEI uses the ontology
alignment API
... does not use SKOS mapping, rather this format allows to
represent the output of matchers, this can then be turned into
SKOS or owl:sameAs mappings
... we need more expressive constructs than owl:sameAs or
owl:equivalentClass
<Fran> http://alignapi.gforge.inria.fr/
oktie: we have been working for a
while (together with IBM) on instance matching / linking.
... something that is missing is metadata/provenance about
terms
... if you look at all HCLS datasets in LOD cloud,
... datasets are highly interlinked, e.g. because they are
based on shared identifiers
... i am not sure how interesting they are for your
initiative
... oktie: i am not sure what the expected input and output
is
francois: we are looking for
datasets that already have links
... last year we used LinkedMDB
<mscottm> Francois: We would like a recommendation of two datasets that are interesting to use as part of the alignment challenge
<Fran> http://oaei.ontologymatching.org/2009/instances/
scott: there are challenges. it is already hard if you have unification of entities, without unifications it gets even harder, but that is the challenge
francois: there are semiautomatic tools like linker or SILK, with further growth of the web of data we need tools with more automation
<mscottm> Scott: So you would like a recommendation of some data sets to use in the challenge?
<mscottm> Francois: Yes.
<mscottm> Oktie: I would guess that a good starting place is the table at http://esw.w3.org/HCLSIG/LODD/Data
oh, i dropped out, although i am still dialed in.
it seems it happened to others as well?
<Fran> yep
<Stephen> i just lost all audio too
<bbalsa> Info about interlinking in LODD, http://esw.w3.org/HCLSIG/LODD/Interlinking
<mscottm> please dial in again if you lost your connection
<mscottm> Scott: Summarizing Francois's request for data sets that have lots of instance data
<mscottm> Scott: Data Updates - nothing new
oktie: nothing new regarding data updates.
matthias: nothing from my side as well.
<mscottm> http://esw.w3.org/HCLSIG/LODD/Mapping_Experimental_Data
scott: if you want to share
datasets, people need information ABOUT the data, not just the
data itself
... how can you support relational systems?
<mscottm> http://sharednames.org
scott: people would also like to
know, for example, who generated annotations, and how.
... there is ongoing work in two other HCLS task forces to talk
about provenance in the context of scientific discourse.
<mscottm> SWAN/SIOC
scott: part of the provenance
group is also working on microarray data, e.g., to find out
about experimental conditions and analytical methods
... in the context of linked open data, you would like to know
how, for example, a dataset was mapped to RDF, a reference to
the set of mappings that were used etc.
<mscottm> Leveraging the Semantic Web and Natural Language Processing to Enhance Drug-mechanism Knowledge in Drug Product Labels
<mscottm> Richard: Leveraging the Semantic Web and Natural Language Processing to Enhance Drug-mechanism Knowledge in Drug Product Labels
i'm back
<scribe> scribenick: matthias_samwald
<rboyce> http://www.pitt.edu/~rdb20/data/DIKB-lightning-summary-05262010-1.pdf
i get a 404 as well
richard: the drug interaction KB
has info from drug labels, in-vitro experiments, FDA guidance
statements
... what drugs inhibit metabolism, et cetera
... we have around 100 drugs in the systems, extensive review
for around 60 of them
... more detail than Drugbank
... we want to discuss possibility this resource as a node in
the LODD cloud
... we are preparing an abstract for a paper in which we want
to create a mash-up using semantic resources
scott: do you already have an idea which ontological terms would be relevant
richard: active moieties / ingredients, biological discourse / evidence models
bosse: is your work not also relevant to TMO?
<mscottm> TMO http://bioportal.bioontology.org/visualize/42758
<rboyce> DIKB: http://www.pitt.edu/~rdb20/data/DIKB-lightning-summary-05262010.pdf
richard: we had problems with
accessing LODD server (at berlin server), we are mirroring
now
... i will also contact anja jentzsch
<Stephen> Have to disconnect now. Thanks all.
<mscottm> thanks!
<Fran> bye
i can't view http://www.w3.org/2010/05/26-hcls-minutes.html
sigh
This is scribe.perl Revision: 1.135 of Date: 2009/03/02 03:52:20 Check for newer version at http://dev.w3.org/cvsweb/~checkout~/2002/scribe/ Guessing input format: RRSAgent_Text_Format (score 1.00) Found ScribeNick: matthias_samwald Found ScribeNick: matthias_samwald Inferring Scribes: matthias_samwald WARNING: No "Topic:" lines found. WARNING: No "Present: ... " found! Possibly Present: DIKB Fran Francois Oktie OktieH P25 P4 P5 Richard Stephen aaaa aabb aacc aadd aaee aaff aagg aahh bbalsa be bosse elgar epichler ericP longer matthias matthias_samwald mscottm now rboyce scott scribenick should trackbot You can indicate people for the Present list like this: <dbooth> Present: dbooth jonathan mary <dbooth> Present+ amy WARNING: No meeting title found! You should specify the meeting title like this: <dbooth> Meeting: Weekly Baking Club Meeting WARNING: No meeting chair found! You should specify the meeting chair like this: <dbooth> Chair: dbooth Got date from IRC log name: 26 May 2010 Guessing minutes URL: http://www.w3.org/2010/05/26-hcls-minutes.html People with action items: WARNING: No "Topic: ..." lines found! Resulting HTML may have an empty (invalid) <ol>...</ol>. Explanation: "Topic: ..." lines are used to indicate the start of new discussion topics or agenda items, such as: <dbooth> Topic: Review of Amy's report[End of scribe.perl diagnostic output]