Date: 24 May 1995 ----------------- ====================================================================== TOPICS: 1. CFP Special Issue on Lexical Choice in TG and MT - MT Journal 2. Book Announcement: A CONNECTIONIST LANGUAGE GENERATOR - Nigel Ward 3. NLP Pacific-rim Symposium '95: 2nd CFP Deadline June 20 1995 ====================================================================== ====================================================================== Topic 1: From: Leo Wanner Subject: CFP Special Issue on Lexical Choice in TG and MT - Machine Translation Journal ******** CALL FOR SUBMISSIONS ******** Special Issue on Lexical Choice in Text Generation and Machine Translation The Machine Translation Journal Editor: Sergei Nirenburg Guest Editor: Leo Wanner (Waterloo and Stuttgart) The Journal of Machine Translation is inviting submissions for a Special Issue on Lexical Choice in Text Generation and Machine Translation. This is the second special issue of the MT Journal that explicitly addresses Text Generation. The first issue (edited by Richard Kittredge, Montreal) was devoted to research in Text Generation that has been applied to Machine Translation or to multilingual language generation. This Issue is devoted to a specific problem --- that of lexical choice --- which arises in both Text Generation and Machine Translation regardless of the underlying theoretical model and the application area of the research. Although lexical choice is of great importance to both Text Generation and Machine Translation, it has very often been ignored as semantic constructions have tended to be associated directly with lexical units. Only recently has lexical choice started to become one of the major areas of research in Generation, and it still remains a significant unresolved issue in Machine Translation. The problem of lexical choice is particularly difficult because it is inseparably intertwined with the problems of syntactic realization, discourse generation, knowledge representation, lexicon organization, etc. This means that solutions to lexical choice require that other tasks in generation and machine translation have already been solved to some degree. This Issue addresses all critical topics in the problem of lexical choice in Text Generation and Machine Translation. One of the basic topics is the placement of the lexical choice process in the generation and translation process, and, subsequently, the range of phenomena that the lexical choice process has to deal with. Another basic topic is the acquisition of constraints for the guidance of the lexical choice process. Until now, lexical choice processes have only rarely "consciously" decided upon the selection of one out of several possible lexicalizations --- although it is generally accepted that decisions are to be made in accordance with (target) language- constraints and culture-specific constraints. An additional basic topic is the most suitable organization of semantic and lexical resources for lexical choice. High quality research papers are invited on these and other topics, including but not limited to: - Finding appropriate lexical equivalents in Machine Translation. - The role of lexical choice in Multilingual Generation. - Acquisition of lexical resources relevant to lexical choice. - Organization of semantic resources and the lexicon for lexical choice. - Design and implementation of lexical choice processes. - Interdependency between lexical choice and the other tasks in generation and machine translation. - Phenomena to be addressed by lexical choice. - Contextual criteria for lexical choice. FORMAT: Articles may be submitted in hard-copy or electronic (plain ASCII or .ps) format to the guest editor. If submitting hard-copy, four copies of the paper are required. The length of the papers should be approximately 20-30 pages (12-point font). DEADLINE: Submissions are due on September, 15 1995 GUEST EDITOR: Leo Wanner Computer Science Department University of Waterloo Waterloo, ON Canada N2L 3G1 Email: lwanner@after.logos.uwaterloo.ca Fax: ++1/ 519 - 885 - 1208 Phone: ++1/ 519 - 888 - 4567 5344 GUEST EDITORIAL BOARD: Robert Dale (Microsoft) Koenraad De Smedt (Bergen) Bonnie Dorr (Maryland) Helmut Horacek (Bielefeld) Richard Kittredge (Montreal) David McDonald (Brandeis) James Pustejovsky (Brandeis) Frank Smadja (Columbia) ====================================================================== Topic 2: From: nigel@sanpo.t.u-tokyo.ac.jp Subject: book announcement nigel Ward nigel@sanpo.t.u-tokyo.ac.jp http://www.sanpo.t.u-tokyo.ac.jp/people/nigel.html Mechano-Informatics, Engineering, University of Tokyo 7-3-1 Hongo, Bunkyo-ku, Tokyo 113 Japan ============================================================================= announcing the publication of A CONNECTIONIST LANGUAGE GENERATOR Nigel Ward Ablex, 1994, ISBN 0-89391-974-8 (hardcover), 1-56750-038-2 (paperback) Based on my 1991 Berkeley thesis, improved in content and exposition, and with a new chapter presenting an alternative to Fillmorean deep case. Abstract This book addresses the task of generating natural language utterances. It is motivated by two difficulties in scaling-up existing generators. Current generators accept only inputs which are relatively poor in information, such as feature structures or lists of propositions; they are unable to deal with inputs containing more information. Current generators also have very restricted knowledge of language; they are unable to deal with more syntactic and lexical options due to the interactions among the possible choices. To address these and other issues I have built a system called `FIG,' short for `Flexible Incremental Generator.' FIG is based on a single associative network which encodes lexical knowledge, syntactic knowledge, and world knowledge. Computation is done by the spreading of activation across the network, supplemented by a small amount of symbolic processing. Thus FIG is a `spreading activation' or `structured connectionist' system. In the initial state, some nodes representing concepts are sources of activation; this represents the input. Activation flows from these nodes to nodes representing words via the various knowledge structures of the network. When the network settles, the most highly activated word is selected and emitted. After this, activation levels are updated to represent the new current state. This process of settle, emit, and update repeats until all of the input has been conveyed. An utterance is simply the result of successive word choices. The treatment of syntax in connectionist and spreading activation systems is a well-known problem. In FIG syntactic knowledge is encoded with network structures representing constructions and their constituents. Constituents are linked to words, syntactic categories, relations, and other constructions. Activation flow via these links, and eventually to words, provides for constituency and subcategorization. The links to constituents are gated by `cursors,' which are updated over time, based on feedback from the words output. This mechanism ensures that words and concepts which are syntactically appropriate become highly activated at the right time; which causes words and concepts to appear in the right order. FIG demonstrates that the complexity present in most treatments of syntax is unnecessary: FIG dispenses with the assembly of syntactic structures; constructions affect the utterance only by the activation they transmit, directly or indirectly, to words. FIG does without a mechanism for explicit syntactic choice; any number of constructions are potentially active, competing or cooperating, in parallel, and the choice among them is emergent. Phenomena traditionally considered to require instantiation and variable binding are handled in FIG with much simpler mechanisms. FIG's syntactic coverage is much broader than that of other connectionist generators; outputs include "once upon a time there lived an old man and an old woman", "one day the old woman went to a stream to wash clothes", and "John ate a peach with an old woman 's fork". FIG is an advance over previous generators: It handles arbitrarily large inputs, since the number of nodes activated in the initial state makes no difference to its operation. It handles trade-offs among competing goals without additional mechanism, since all computation is in terms of numbers. FIG also handles interaction among choices easily, since it tends to settle into a state representing a compatible set of choices, due to links among nodes representing such choices. Abstracting from the implementation leads to several design principles for generation, including explicit representation of the current state and many uses of parallelism: `knowledge source parallelism,' the simultaneous activity of syntactic, lexical, and other considerations; `competitive parallelism,' the simultaneous activity of multitudes of alternative words and constructions; `part-wise parallelism,' the simultaneous consideration of words for all parts of the utterance; `evaluative parallelism,' the simultaneous evaluation of all sources of evidence for the relevance of a word, construction, or concept; and `synergistic parallelism,' the simultaneous activity of many constructions. Corroboration for these design principles is found in the fact that human speakers appear to manifest them, as evidenced by introspection and data on pauses, priming, and speech errors. FIG also points the way to more natural machine translation. In existing generators the structure of the input closely governs that of the output, but FIG is more flexible: there is no direct mapping from its input to its output, which makes it better able to take advantage of idiosyncrasies of the target language. ====================================================================== Topic 3: From: nlprs95 Subject: NLPRS '95: 2nd CFP +----------------------------------------+ | NLPRS '95 - CALL FOR PAPER | | (second announcement) | | | | THIRD NATURAL LANGAUAGE PROCESSING | | PACIFIC-RIM SYMPOSIUM | | | | Seoul, Korea | | December 4-6, 1995 | +----------------------------------------+ NLPRS'95, which will be held from December 4-6, 1995 in Seoul, Korea, is the third forums to bring together NLP researchers in the Pacific-Rim area for scientific exchange and presentation. The program will include tutorials, invited talks, and demonstrations as well as tracks for paper and video presentations. TOPICS OF INTEREST ~~~~~~~~~~~~~~~~~~ Submissions are solicited on original and previously unpublished research in all aspects of NLP, including, but not limited to: * phonetics * generation * phonology * parsing * morphology * machine(-aided)translation * lexicon * spoken language processing * syntax * linguistic models of natural language * semantics * natural language interface and dialog systems * pragmatics * language-oriented information retrieval * discourse * corpus-based language modelling * document analysis * multimedia and language INVITED LECTURERS ~~~~~~~~~~~~~~~~~ The following researchers have already confirmed their paticipation in NLPRS as guest speakers: Eugene Charniak, Brown University (USA) Kenneth Ward Church, AT&T (USA) Sadaoki Furui, NTT (JAPAN) Changning Huang, Tsinghua University (CHINA) PAPER SUBMISSION ~~~~~~~~~~~~~~~~ Papers of no longer than 6 pages in the double-column conference format should be submitted by 20th June, 1995. We strongly encourage papers to be electronically submitted. In this case, they should be in LaTeX format, plain text or PostScript format for non-alphabet, and should be emailed to : nlprs95@cair.kaist.ac.kr Latex submissions must use the NLPRS submission style (nlprs.sty) retrievable from the NLPRS ftp server or WWW Home Page. The following is an example of getting the NLPRS submission style sheet by anonymous FTP: $ftp cair-archive.kaist.ac.kr Name:anonymous Password: ftp> cd NLPRS-95 ftp> get nlprs.sty ftp> quit If electronic submission is not possible, three hard copies of the paper should be sent to: Mr. J.M. KIM NLPRS'95 Secretariat Foreign Tourist Dept II Hanjin Travel Service Co.,Ltd.(Conference Agency) 132-4, 1-ka, Bongrae-dong, Chung-ku, 100-161, Seoul, Korea Phone:+82-2-726-5540, Fax:+82-2-773-1623 General Chair ~~~~~~~~~~~~~ Gil-Chang Kim Korea Advanced Institute of Science and Technology International Advisory Committee ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Changning Huang (Tsinghua University) Young-Taek Kim (Seoul National University) Makoto Nagao (Kyoto University) Hirosato Nomura (Kyushu Institute of Technology) Vilas Wuwongse (Asian Institute of Technology) Tae-Ok Kim(Sogang University) Chung-min Lee (Seoul National University) Program Committee ~~~~~~~~~~~~~~~~~ Chair: Hozumi Tanaka (Tokyo Institute of Technology) Vice-Chairs: Key-Sun Choi (KAIST), Changning Huang (Tsinghua University), Hitoshi Iida (ATR) Members: Terumasa Ehara (NHK), Seong-Guk Han (Wonkang Univ.), Young-Gyun Han (Univ. of Ulsan), Koichi Hashida (Electrotechnical Laboratory), Satoru Ikehara (NTT), Gunbae Lee (Postech), Hsi-Jian Lee (National Chiao Tung Univ.), Jong-Hyeok Lee (Postech), Hae-Chang Lim (Korea Univ.), Hiroshi Maruyama (IBM Japan Ltd.), Yuji Matsumoto (Advanced Institute of Science and Technology), Kazunori Muraki (NEC Corporation), Yoshihiko Nitta (Hitachi Ltd.), Jonathan Oh (Univ. of Missouri at Kansas City), Seyoung Park (ETRI), Dong-Yul Ra (Yonsei Univ.), Jung-Yun Seo (Sogang Univ.), Chew Lim Tan (National Univ. of Singapore), Lua Kim Teng (National Univ. of Singapore), Takenobu Tokunaga (Tokyo Institute of Technology), Benjamin K. T'sou (City Polytechnic of Hong Kong), OM Vidas (Embassy of India, Tokyo), Shiwen Yu (Peking Univ.), Zaharin Yusoff (Univ. Sains Malaysia), Ming Zhou (Tsinghua Univ.) Organizing Committee ~~~~~~~~~~~~~~~~~~~~ Chair: Key-Sun Choi (KAIST) Members: Dong-Un An (Chonbuk National Univ.), Hee-Rahk Chae (Hankuk Univ. of Foreign Studies), Jin-Hee Choi (KAIST), Young-Suk Han (KAIST), Dosam Hwang (SERI), Seung-Shik Kang (Hansung Univ.), Nam-Kyeng Kim (KAIST), Myoung-Wan Koo (Korea Telecom.), Hyuk-Chul Kwon (Pusan National Univ.), Kang-Hyuk Lee (Korea R&D Information Center), Yong-Ju Lee (Wonkang Univ.), Yong-Seok Lee (Chonbuk National Univ.), Hyung-Nam No (Korea Univ.), Yongkyoon No (SERI), Dong-In Park (SERI), Jae-Deuk Park (SERI), Gi-Chul Yang (Mokpo University), Jae-Woo Yang (ETRI) DEADLINES ~~~~~~~~~ Paper Submission: June 20, 1995 Notification of Acceptance: August 30, 1995 Camera Ready Copy Due: September 30, 1995 ACCOMMODATIONS ~~~~~~~~~~~~~~ Hotel Sofitel Ambassador is designated for the Symposium accommodation. Room rate is specially reduced as US$100.-/night approximately for single, double or twin type. The above rate includes 10% service charge & 10% VAT, but it does not include breakfast. Hotel reservation should be made through the Secretariat. PLANNING TO ATTEND ~~~~~~~~~~~~~~~~~~ Researchers planning to submit a paper or/and to attend the NLPRS meeting are asked to complete and return the interest form below by fax or e-mail to NLPRS '95 Secretariat. It will help the conference organizers estimate the facilities needed for the conference and will enable us to provide all interested people with updated information. +----------------------------------------------------------------+ | REGISTRATION OF INTEREST | | | | Prof/Dr/Mr/Ms . . . . . . . . .. . . . . . . . . . . . . . . . | | Affiliation . . . . . . . . . . . . . . . . . . . . . . . . . | | Position . . . . . . . . . . . . . . . . . . . . . . . . . . . | | Postal Address . . . . . . . . . . . . . . . . . . . . . . . . | | . . . . . . . . . . . . . . . . . . . . . . . . | | Country . . . . . . . . . . . . . . . . . . . . . . . . . . . | | Telephone. . . . . . . . . . . . . . Fax . . . . . . . . . . . | | Email address. . . . . . . . . . . . . . . . . . . . . . . . . | | I intend to submit a paper (yes/no). . . . . . . . . . . . . . | | I expect to attend NLPRS '95 (yes/no). . . . . . . . . . . . . | +----------------------------------------------------------------+ ****************************************************** * The WWW version of this CFP is also avaiable at: * * * * URL:http//cair.kaist.ac.kr/~nlprs95/NLPRS95.html * ****************************************************** ====================================================================== eof ======================================================================