WINTER SCHOOL Report, July 2006 Steven Bird ACL/HCSNet Advanced Program in Natural Language Processing University of Melbourne, 10-14 July 2006 http://lt.csse.unimelb.edu.au/nlp06/ The HCSNet/ACL Advanced Program in Natural Language Processing consisted of a series of advanced tutorials on major topics of current research interest in NLP. In developing the program our goals were twofold: to encourage international students to come to Australia for the COLING/ACL by providing a second week of NLP content, and to give Australian students the benefit of extended interaction with researchers who will be attending the COLING/ACL. We were delighted to welcome 70 participants from 15 countries. PROGRAM The program was organised by Nicola Stokes, Lawrence Cavedon, Timothy Baldwin, and James Curran. The NLP/IR and Discourse & Dialogue tracks ran in parallel in the mornings, while the Lexical Semantics and Probabilistic Parsing tracks ran in parallel in the afternoons. Full abstracts and bio sketches were posted on the website. NLP/IR: Sophia Ananiadou (Manchester), Francis Bond (NTT Japan), Inderjeet Mani (MITRE and Brandeis), Alistair Moffat (Melbourne) DISCOURSE AND DIALOGUE: Ivana Kruijff-Korbayova (Saarland), Diane Litman (Pittsburgh), Michael Strube (EML Research), David Traum (USC) LEXICAL SEMANTICS: Collin Baker (Berkeley), Timothy Baldwin (Melbourne), Paul Buitelaar (DFKI), James Curran (Sydney), Graeme Hirst (Toronto) PROBABILISTIC PARSING: Stephen Clark (Oxford), James Curran (Sydney), Steven Bird (Melbourne), Anoop Sarkar (Simon Fraser) SPONSORSHIP AND BUDGET The ACL provided A$20k sponsorship, and this was mainly used to fund 33 scholarships (covering registration and accommodation). The ARC Network in Human Communication Science (HCSNet) also provided A$20k sponsorship, and this was mainly used to fund 14 presenters (covering accommodation and travel subsidy). Other sponsors were the Department of Computer Science and Software Engineering, the University of Melbourne, and National ICT Australia. Budget summary (approximate figures only, pending final accounting) Aus$ Comment +45,000 Income (sponsorship and registrations) -13,000 Student accommodation (218 nights in student dorms) - 6,000 Presenter accommodation (38 nights in hotel) - 9,000 Presenter travel subsidy -10,000 Conference management, administration, materials - 7,000 Catering ---------------------- = 0 ACL SCHOLARSHIPS Applications were evaluated according to the following criteria, listed in order of importance: (i) commitment by student's institution to fund travel to Australia; (ii) whether the student has a paper accepted at the COLING/ACL conference or a satellite event; (iii) student's curriculum vitae; (iv) relevance of school program to student's research project. All applicants who had papers accepted for the COLING/ACL conference or a satellite event were offered scholarships. This exhausted the scholarship funding. Scholarship holders came from the following countries: Australia (8), Brazil (2), Canada (3), China (4), Germany (2), India (2), Ireland (1), Israel (1), Japan (2), Netherlands (1), Singapore (1), Sri Lanka (2), UK (1), USA (3). EVALUATION Summary evaluation: Presenters and participants alike felt that this was a useful event, building on the travel and organisation of the main conference, but permitting more relaxed interaction. In regions where NLP summer schools are not already established, this proved to be an effective way to give graduate students more extended and tailored input than they get from the main conference. Detailed evaluation: 43 participants completed an evaluation form. Of the 25 scholarship holders coming from outside Australia, 14 indicated that the existence of the event helped them get sponsorship to travel to Australia. All 43 respondents said they would recommend this kind of event to other students. Participants rated the program in four areas, scoring it in the range 1-5 (1=bad, 5=good), and providing free text comments: A. Program: 4.2 (quality of content, timetable, topic selection, level) Some participants pointed out the varying levels of difficulty within a single theme; several said that the content was not advanced enough for them; no-one complained about the choice of tracks that were scheduled concurrently. Representative quote: "A very nice overview of each area, with good scheduling". B. Benefit: 4.0 (helpfulness to your research, interaction with presenters) Most participants said the program was very helpful, filling in gaps and giving them more extended access to presenters than the main conference. Representative quotes: "Very beneficial; I learnt a lot, deepened my knowledge and met presenters and had very useful discussions with them, in addition to other students"; "The smaller group before the main conference was an excellent opportunity to meet and mingle. The mix of topics was good. I learnt a lot." C. Local Arrangements: 4.4 (accommodation, catering, venue, registration) Participants praised the venue and social program, but some were critical of the student accommodation (too basic, lacking in internet connectivity). Representative quote: "Well organised event, excellent accommodation, and tasty coffee breaks". D. Organisation: 4.4 (website, communication, scholarship process) Participants were generally very positive, though a few pointed out that the conference management had been slow to confirm registrations, and that the program for one of the tracks was not finalised until a few days before the event.