AKT project meeting, Southampton 25 Jan 2005 Lynda and Lloyd Lynda's general impression - the AKT propject got a lot of funding and the quality of the people working in the project is very high (most are from groups with the highest classification in the UK system). I think the work they are doing shows this. There are 3 or 4 groups that we should definitely keep in touch with (Southampton, KMI, Sheffield). The thing that really struck me this time is that people were stating explicitly that they saw that our work was complementary to what they were doing and seemed to be excited about it. This was very encouraging. Georg Stork also knows who we are - also nice. (The most important thing of all is that Wendy just loved my outfit :-) ) TUESDAY MORNING: DEMOS I (Lynda) had long chat to Nick Gibbins, Steve Harris about their cool ontologies plus Microcosm demo (AKTive Futures). It is known technology put together, but it is really neat. They also combine it with IR techniques. Build an ontology by hand, analyse existing web pages and classify them into the ontology, then add related pages using "classic" IR techniques. Nice example with enhanced BBC news page (ontology on energy sources). They also have display of numeric data: oil production of different countries per year. I pointed them to Thomas Kamps' book and Amit's thesis as a quick source of information on how to display things. mSpace and showed Topia to three guys working with monica. They are going to do a mobile/web service version of mSpace (Microsoft is interested). Hans-Georg Stork, EU Call for e-content - getting Sem Web technology into musea. Call around end 2005. (This could be extremely relevant for Lloyd, working with the Rijksmuseum, to take the work one step further. Not to mention our other connections with DEN etc..) Vita: Vitaveska Lanfranchi http://www.dcs.shef.ac.uk/~vita/ Chatted to Vita about her work (recommended Thomas Kamps book...). Katya should definitely have another chat with her to compare how both their ideas are progressing. (Katya visit Sheffield?) (I found it difficult to explain exactly what the underlying ontologies look like - is this in the current paper?) I (Lloyd) did indeed speak with Hans-Georg just after Lynda did. I will track this. Would be nice to use this to take national CATCH CH repository integration to an international (well, EU) level. I (Lloyd) mostly gave demos. For a funding guy, Hans-Georg was pretty involved with Topia/Noadster. I chatted a lot, both at demos and through meeting, with Harith Alani and Mischa (a male) Tuffield of Southampton. They are both interested in narrative. Harith is from Artequakt, and Mischa is just starting a PhD and is keenly interested and informed in narrative as well. I see a promising collaborative future with these two. Mischa is interested in interactive narrative in a gaming context. His supervisor (Nigel) is not too keen about the "seriousness" of the topic of gaming - so we tried to reassure him that this was both serious and difficult. --- TUESDAY AFTERNOON: WHIRLWIND AKT TOUR The afternoon was a machine gun firing of 20 minute talks from each group, each of which was a machine gun firing of shots from each of that group's projects. Impressive in quality and scale overall, to the degree you can measure with such quick overviews. --- Nigel interested in semantic web enabled personal life capture. (Sem Web "Steve Mann".) AKT 6 year project. Started year 2000. Ends Oct 2006. Book of selected AKT papers. Developing and consolidating infrastructure. Show app benefit. Acquire, model, reuse, retrieve, publish, maintain knowledge. Nice overview of the different sub-projects in AKT. Semantic interoperabilities. Linking heterogeneous sources. (Any chance of access to Nigel's slides?) --- Simon Buckingham Shum, KMI OU http://kmi.open.ac.uk/people/sbs/ Knowledge-intensive distributed collaboration People have base-line tools. Consensual knowledge, contested knowledge orthogonal formalized knowledge, unformalized knowledge Add semantics into the tools at time of use. MM presentation generation (mmsecond representation of time) CoAKTinG toolset - they use off-the-shelf environments. Video taken of a meeting, and annotated - you can navigate the video in terms of the decisions made (for example). papers available They are involved with a NASA project to allow collaboration with Mars scientists, earth scientists and then earth to mars scientists! Meeting replay environment. Paper in eprints archive. http://eprints.aktors.org/ --- Dave Robertson, Edinburgh http://www.dai.ed.ac.uk/groups/ssp/members/dave.htm Mathematics of peer to peer (sorry, not my scene). (Instance calculus.) Dynamic ontology mapping (Fausto's work in Trento) Interaction model gets passed around and takes the mapping commitments with it. --- Sheffield NLP originally. http://www.aktors.org/technologies/armadillo/ --- Enrico Motta, KMI OU http://iswc2004.semanticweb.org/demos/40/ http://kmi.open.ac.uk/projects/magpie/main.html New Magpie architecture --- Aberdeen (Sorry, missed it because doing my (Lynda's) own stuff.) --- Nigel Shadbolt, Southampton User i/fs for rich media semantic spaces Swoogle (Jim Hendler) ranking ontologies http://www.swoogle.org http://pear.cs.umbc.edu/swoogle/ EPSRC application (this is one Nick demoed to me). They couldn't find a common data structure but now everything is converted to RDF. So suddenly they had access to integrated information. They have been working with citeseer. They have a US version of AKTive space. AKTive Futures. TBL is joining Southampton for 20% of his time. --- TBL from the USA On video with a mike for the whole room. (He has the slides from the speakers on the web - where are they? I want them!) Semantic Web should become more Webby. (This is good to hear!) Common API across whole community. Data formats are OK, but protocols [are needed, I guess]. Ontologies and speech. He sees the future are taking an ontology and then generating natural language out of it (at least, if I understood him right). (The sound wasn't good enough to really hear what he was saying.) --- Wed 26 Jan Lloyd (Lynda: Lloyd spent too much time on first half of talk and too little on second half.) For 9:30am, no coffee service and full wireless (not to mention my (Lloyd) aforementioned obsession with opening sheet bullet point text), the audience looked relatively attentive. Gary Wills asked about using museums other than the Rijksmuseum, and I discussed some of the merging stuff we've dabbled with and the domain-independence of Noadster. Also elaborated on how with merges, Noadster's structuring would still work. monica commented with an end-of-talk question that I never used the word "ontology". I discussed how the document structure generated does not necessarily mirror the underlying ontology, but is affected, and how. Perhaps monica just wanted to know that we were using an ontology at all. --- Jim Hendler http://www.cs.umd.edu/~hendler/ He also had a generated URL of an RDF web page (human unreadable) as his Semantic Web ID. Semantic Web "Current Wave" slide. (This was a nice overview slide, but I couldn't track it down :-( ) Research -> wide take-up. Standards World SW tools, Portals, Markup http://www.mindswap.org/ [Can we do anything with this? http://www.mindswap.org/photos/ ] Social networking and SW FOAF is most successful thing on semantic web. Millions of people known to FOAF. A thin ontology with very many instances. Discussion about building gene ontologies. I was trying to sign up realtime to Film Trust - a SW film recommender. http://trust.mindswap.org/FilmTrust/ Collaborative editing http://www.mindswap.org/2004/SWOOP Policy aware web http://www.mindswap.org/~hendler/2004/PAW.html Information lives in specific contexts (the AI people have been saying this for years of course). Sem Web allows us to place the info in (multiple) contexts. --- Lynda chatted to Paul Lewis (international museum SW work - excellent for CATCH/N9c connections), Gary Mills (talk to Les Carr and Tim Miles Board) for Katya. Simon also keen on keeping in contact with KMI. New person doing narrative stuff and Lloyd and/or I invited over by Enrico. --- MIAKT (Pronounced "My Act") Generic KM architecture for Medical Imaging. Document generation services 4000x3000 pixel images Radiologist draws round interesting part of image. http://www.aktors.org/miakt/ --- Georg Stork Knowledge & Cognition in FP6-IST "after the final AKT.... ...EUphoria?!" IST big part of framework programme. IST (Info Soc Technologies) Semantic-based Knowledge and Content Systems Cognitive Systems Call 4 1.1 BEuro, 22 March 2005 projects start Jan 2006 Call 5 0,6 BEuro, to be published, /-May 2005 Sem-based Knowledge and Content Systems (SO 2.4.7. ~100 MEuro) 1. knowledge acquisition and modelling 2. knowledge sharing & use exploit semantics for access, integration, interoperability **3. intelligent content sophisticated architectures for multi-hypermedia content, production, delivery and presentation www.cordis.??? Cognitive Systems, real world, real people, (Ties in nicely with ToKeN2000 idea of CS & Cognitive Sciences) FET Future and Emerging Technologies E-contentplus Non research, but applying stuff in real life contexts (musea stuff?) ------ OFFLINE MEETING We were given a slot and room to meet with people who wanted to met with us. These were monica, David Millard, Mischa, and Harith. We only had 45 minutes, but it was a fun talk in the best academic sense. Discussion included the relation between "skeletal" narrative ala mSpace and Noadster vs the deeper narrative ala DISC and Artequakt. Also included the spread between path-based narrative vs narrative emerging from rule-based rules and character agents. Mischa's background is with such agents and their emerging narrative. We may to another trip to Southampton in February or March to meet with them again and project partners from Sheffield. ------------------------- (New TLA - DTS - Doctoral Training Student)