wed 5 may - tue 11 may algorithm week -> adaptive IE algorithms, meaning: elie (convergent boundary classification), bien (dynamic bayesian network), lp2 (symbolic rules & lightweight nlp), rapier (inductive logic programming), whisk (top-down rule extraction), srv (multi-strategy, with two learners and a naive bayesian classifier), bwi (wrapper patterns and boosting), hmm (well.. hidden markov models) and other classics (including brill's pos tagger) tending to favour the nlp approach for IE over the statistical approach since it -inherently- enables an easy introduction of the cidoc crm ontology into the IE process -- tools to be tested for potential use & integration following this approach: gate (link), ellogon (link), ontomat (link), melita (link) -- tools to be tested for potential use & integration following an alternative approach: fastr (link), intex (link) impatiently waiting for.. 1. a public release of the javanlp (link) and ontolt (developed within sig5 of ontoweb) plug-ins to protege (link) 2. mes potes (may 21st is just around the corner :) 8:45:28 PM

wed 28 april - tue 4 may this was a home week -- wandering through information extraction bibliography, debugging ie tools.. not too much to write about though (it took me a considerable amount of effort to get kaon, cohse and mnm to work properly & get accustomed to the tomcat servlet and the java driver for mysql) cool things during home week: ordered some excellent books for work (link & link) and play (link), saw norah (link), got my tkts for nuits sonores 2nd ed. (link) uncool things during home week: software f...-ups, op-art not funded (link), asse-angers 0-0 7:49:25 PM

tue 27 april spent the whole day studying document architecture literature -argumentative zoning / rhetorical structure theory- .. topic maps in xtm (link) or rdf (link) syntax could be used to express document structure semantics semi-automatically (namely, without having to manually annotate every single document) -- the idea being the following:  a. create a pool of topic maps -a taxonomy of cultural heritage document structures- using a basic corpus of texts and  b. the learning process: compare the structure of every document to this pool in order to determine what kind of document it is -- hence, we would get a good idea of what to expect from this document not only in terms of structure (where to find the abstract / the references / the quotes) but also in terms of writing style and other semantic riches maybe this will work -- the thing is, it has to work with documents distributed over a HyperCuP-topology-structured Internet space (link), in conjuction with a formal cultural heritage ontology (link) used to express content semantics in those same documents :| .. it's been nice knowing you! 10:43:41 PM

mon 26 april well well.. it was a great MONTH indeed :D -- i'm back in lyon since last wednesday, first day @the office though -- events the World should know about: cadabra code (link) have a garageband site (link) with some stuff to read & listen to, huge win over paok yesterday (1-2, olisadebe scoring the winner in stoppage time), cyprus said no to the final annan plan (link) & joins the EU next week excellent news from the caa 2004 conference (link) including an opportunity to apply the framework discussed in our caa 2003 paper (peer-to-peer ways to cultural heritage) in a pilot project -- more on this later.. research work: currently looking into the semantic analysis of document structural markup and surveying text mining state-of-the-art (clustering, classification, retrieval) -- together with peer-to-peer computing & ontological engineering state-of-the-art, this work will shape up into the final version of my first two thesis chapters 8:49:26 PM

fri 26 march packing up for bxl -- leaving tomorrow @4pm on the tgv -- looking forward to a great week of work & play, crowned by a w-e in amsterdam (link) 6:02:53 PM     fri 19 march - thu 25 march spent the whole week preparing my thesis intermediate presentation -- the event took place at ERIC, where the supervising committee convened today -- everything went down very well, and a schedule for the remaining research work was discussed and agreed upon the big M is september 2005 :) 5:59:02 PM

thu 18 march edutella hypercubes is the order of the day -- plus ERIC laboratory photo shoot for next year's Lyon 2 research guide * the gathering play @rail theatre tonight -- 2nd time i will be seeing them live -- 1st time @london astoria 2 with pol, giorgos and gabriel ** huge away victory (2-3 sets) in the 1st game of the volleyball championship finals 1:35:55 PM     wed 17 march worked from home today -- not much though -- got a little bit philosophical about the whole Saint-Patrick's Day thing ;) 1:30:26 PM     tue 16 march a bit of jxme (link) work today, just to test the functionality of all available tutorials for my final year student -- went a little bit further, experimenting with various exchanges between a midp chat client, a chat client implemented through the shell (link), and a webstarted myjxta2.2a (link) 1:27:51 PM

mon 15 march edutellaedutellaedutellaedutellaedutella -- you get the idea... 3:59:07 PM     sun 14 march day off! :) asse-caen au chaudron (link), wonderful weather and a very nice lunch with friends -- topped up by a 0-1 result away to nea smyrni while osfp lost at home to aek -- what more can you ask for ?? 3:58:06 PM     sat 13 march wiki development work part 2: adapted the phpwirelessmap tool (link) by the nantes wireless team (link) to suit our wiki platform needs -- new code available from here soon -- many thanks to ludovic toinel (email) for making this contribution to the open source community 3:53:26 PM     fri 12 march code browsing continues... with wiki development interludes for our caa 2004 demo presentation (FCS_WORD) -- the museum of byzantine culture (link) and the natural history museum of the lesvos petrified forest (link) have already agreed to commit information to the new platform 3:43:52 PM

© Copyright 2004 Efthimios Mavrikas.