HCLSIG/LODD/Meetings/2009-09-16 Conference Call
Conference Details
- Date of Call: Wednesday September 16, 2009
- Time of Call: 11:00am Eastern Daylight Time (EDT), 16:00 British Summer Time (BST), 17:00 Central European Time (CET)
- Dial-In #: +1.617.761.6200 (Cambridge, MA)
- Dial-In #: +33.4.89.06.34.99 (Nice, France)
- Dial-In #: +44.117.370.6152 (Bristol, UK)
- Participant Access Code: 4257 ("HCLS").
- IRC Channel: irc.w3.org port 6665 channel #HCLS (see W3C IRC page for details, or see Web IRC)
- Duration: ~1h
- Convener: Susie
Agenda
- iTriplify Outcome - Anja
- Application demo - Jun
- Web questions demo - Anja
- Paper/conference opportunities (AAAI Spring Symposium on Linked Data Meets Artificial Intelligence, BMC TCM, ISMB, Semantic Web Challenge) - all
- Data update - Anja, Jun, Matthias
- Open data - Susie
- F2F - Susie
- AOB
Minutes
Attendees: Julia, Oktie, Anja, Bosse, Susie, Matthias, Scott
Apologies: Kei
<Julia_Kozlovsky> Anja reports on the First Prize
<Julia_Kozlovsky> Anja: second prize went to Austria and third to Germany
<matthias_samwald> link to the screencast.
<matthias_samwald> http://www4.wiwiss.fu-berlin.de/lodd/triplification_challenge/talk.html
<Julia_Kozlovsky> Anja: also got 1000 Euro check
<Julia_Kozlovsky> Anja: the win was probably due to having best use cases
<Julia_Kozlovsky> Susie: can the money be spend to enable travel for F2F?
<Julia_Kozlovsky> Oktie: Anja should take what's needed to cover her expenses and work..
<Julia_Kozlovsky> Anja: let's get the money first and then decide
<Julia_Kozlovsky> Susie: who will come to the F2F in CA? This will help determine what to do..
<Julia_Kozlovsky> Oktie: right before a deadline, so might come by for a little bit
<Julia_Kozlovsky> Anja: may be a bit expensive for a two-day meeting..
<Julia_Kozlovsky> Susie: there may be options for virtual meet-up
<Julia_Kozlovsky> Susie: there also may be a conflict with other tasks (ex. Translational Medicine Ontology) during a two-day F2F
<Julia_Kozlovsky> Bosse: still a maybe, but working on it
<Julia_Kozlovsky> Matthias: no, can't make it..
<Julia_Kozlovsky> Susie: what should the F2F focus on? Does it make sense to combine with the Translational Ontology group?
<Julia_Kozlovsky> Consensus: yes, it makes sense
<matthias_samwald> there could be another opportunity for a meetup: there will be a "Linked Data Camp" in Vienna on December 1-2. This is before ESTC, so some people might be there anyways. See http://www.meetup.com/Linked-Data-Camp/de/calendar/11318241/
<matthias_samwald> (maybe someone could mention that on audio :) )
<Julia_Kozlovsky> Susie: TMO is working on finishing the ontology, getting use cases and application pilot to submit to Journal of Biomedical Informatics
<Julia_Kozlovsky> Susie: if the Journal extends the deadline, the TMO group might be too busy finishing their work to focus on anything else..
<Julia_Kozlovsky> Susie: a natural focus for both groups might be use cases..
<Julia_Kozlovsky> Julia: what about working out some technical details?
<Julia_Kozlovsky> Susie: we still need to get familiar with each other's work on top level, before talking details..
<Julia_Kozlovsky> Susie: one possibility is for the TMO group to use some of Jun's application code?
<Julia_Kozlovsky> Anja: not sure if the use cases match..
<Julia_Kozlovsky> Susie: TMO hasn't yet picked out their use case, so one factor may be the ability to re-use Jun's software (other factors are availability of data and so on)
<Julia_Kozlovsky> Anja: we should also talk about some visualization ideas later today and tie that back to the application discussion
<Julia_Kozlovsky> Susie: will email Jun
<anjeve> http://www4.wiwiss.fu-berlin.de/lodd/topquestions/
<Julia_Kozlovsky> Anja: outputs may be grouped by sources of questions (pharma vs. clinicians)
<Julia_Kozlovsky> Anja: hover over dataset to see links to other sets, questions on the sets and objects in the data sets
<anjeve> http://dbpedia.neofonie.de/browse/
<Julia_Kozlovsky> Anja: above is example of faceted browsers on top of SPARQL endpoints
<Julia_Kozlovsky> Anja: for our data, the users can see things like ratings: which targets are most linked to from drugs and so on..
<anjeve> http://www.freebase.com/labs/parallax/
<Julia_Kozlovsky> Anja: another approach is by Freebase: another faceted browser, but a bit slow..
<matthias_samwald> faceted browsing of aTags (including SIDER): http://hcls.deri.org/atag/explorer/
<Julia_Kozlovsky> Anja: since the data is already available and linked, users may be presented with a choice of browsers?
<matthias_samwald> try "Varenicline". Does not work in IE.
<anjeve> http://relfinder.dbpedia.org/app.swf
<Julia_Kozlovsky> Anja: Relation Finder offered to adjust their app to our needs/data
<Julia_Kozlovsky> Anja: if we have confidence values for links, this can be used in visualization as well
<Julia_Kozlovsky> Anja: annotated links can also be used to determine relevance
<Julia_Kozlovsky> Anja: people need to know how concepts are linked, define their own questions, transform natual language queries to SPARQL
<Julia_Kozlovsky> Anja: we can define natural language to SPARQL mappings if we have a stable ontology
<Julia_Kozlovsky> Anja: again, different types of users might need different interfaces (NLP tools vs. faceted browsers)
<Julia_Kozlovsky> Susie: the application UI is limited, so a faceted browser and relation finder are important in the long run, but the "Top Questions" approach works well as a pilot
<Julia_Kozlovsky> Anja: the Relation Finder creators (from Germany - it's open code) are willing to work with us if we can provide a use case..
<Julia_Kozlovsky> Anja: the Relation Finder would work better with confidence scores. Working on how to implement this with triples..
<Julia_Kozlovsky> Susie: all background on links is important
<Julia_Kozlovsky> Anja: we can also include the info on the linker itself for people to decide whether to trust the link
<Julia_Kozlovsky> Julia: how about a visualization topic for the F2F?
<Julia_Kozlovsky> Anja: yes, we can look at the browsers and the Relation Finder, hopefully with a use case..
<Julia_Kozlovsky> Susie: good idea, will have to look at timing..
<Julia_Kozlovsky> Anja: there is a possibility of getting an FDA data set, but not sure who the person was..
<Julia_Kozlovsky> Susie: Oktie might know - will email
<Julia_Kozlovsky> Susie: let's see if the questions can be answered through public data sources in the meantime
<Julia_Kozlovsky> Anja: where does approval data come from? Also, what do patients say about drugs?
<Julia_Kozlovsky> Anja: what datasets have patient feedback?
<Julia_Kozlovsky> Anja: also, datasets on lifestyle changes, and patents on pathways and targets..
<Julia_Kozlovsky> Susie: let's get a list of potential data sets based on un-answered questions..
<Julia_Kozlovsky> Matthias: how about an email discussion?
<Julia_Kozlovsky> Anja: will email the HCLS group..
<Julia_Kozlovsky> Anja: need details on the Translational Medicine Ontology - will discuss tomorrow
<matthias_samwald> yes!
<Julia_Kozlovsky> Julia: for now, we are using the "first reasonable link" approach in TMO..
<Julia_Kozlovsky> Anja: any ideas of which conferences to submit the bigger version of the paper?
<Julia_Kozlovsky> Anja: can talk about this next time..
<Julia_Kozlovsky> Anja: an issue had been raised about the rights to all data sets, but all the data we use is free for researchers. Let's discuss more at the next meeting..
<Julia_Kozlovsky> Julia: should there be an organized HCLS effort to make sure we have the rights to data and/or apps..
<Julia_Kozlovsky> Anja: let's wrap up.