Date: Tuesday, October 1, 1996 (preceding the AMTA-96 conference)
Place: Montreal, Canada
The SIG-IL of the AMTA is organizing a ``pre-workshop'' workshop to be held the day before the opening of the AMTA-96. The purpose of the gathering is to bring together those in the research community who have been involved in the development of interlinguas and interlingual approaches to MT (1) to review the state of the art, (2) to identify a range of open theoretical and applied issues related to the development and use of interlinguas in machine translation and other applications, and (3) to select those issues which are most pressing and plan a more extensive workshop for 1997 during which these issues will be addressed more thoroughly.
The pre-workshop workshop will itself be informal and its proposed structure reflects this. Below we list a number of general topics and request that those who wish to participate indicate:
which topic they would like to summarize in term of the current state of the art,
or which issues within a given topic they would like to address, stating VERY BRIEFLY but CLEARLY what the nature of their position on each issue is.
The responses are expected to be between 2 and 5 pages (500 and 1200 words) in length but there are no actual requirements beyond being clear and concise. The issues are expected to emerge from the responses; thus, only topics will be mentioned below.
From the responses, members of the AMTA SIG-IL committee will develop a number of panels (four to six) to present the state of the art with respect to a topic and to present positions and lead discussions about issues related to that topic. As much time will be left unscheduled as scheduled for ample discussions of the issues raised. Position papers should focus mainly on clarifying the issues and stating substantive positions. Novel but underdeveloped positions are encouraged so long as they cause people to rethink standard positions or see old issues in a new light.
Topics:
What is an interlingua?
What information is captured by an adequate interlingual
representation system?
How can interlingual representation systems be built or scaled
up?
How can interlingual representation systems be evaluated?
What can interlingual representations be used for?
They first topic concerns what interlinguas are from one or another perspective (functional, theoretical). The second concerns the sorts of information that need to be, or are likely to be, represented by interlinguas and how that information should be represented. The third concerns how interlingual representation systems might be built (operational criteria for deciding what to include or not to include, the knowledge sources needed to support the construction of IL representations, mechanisms for facilitating the knowledge process, methods for scaling interlingual systems up from a demonstration level to covering larger, possibly unrestricted, domains. The fourth concerns how IL representation systems might be evaluated or compared. The fifth concerns how, apart from their role in support of machine translation, interlinguas might be applied to various other information processing tasks (e.g., text summarization, information extraction, query systems, information retrieval, tutoring, multimodal communication, and the like).
Participation will be limited to 30, mainly selected on the basis of reviews of the responses submitted in conjunction with reviews of vitae. A small registration fee of not more than $10 may be requested to cover administrative costs.
Please send responses to (email preferred, but fax acceptable):
David Farwell
Computing Research Laboratory
New Mexico State University
Box 30001 / 3CRL
Las Cruces, NM 88003
fax: 505-646-6218
email: david@crl.nmsu.edu
Please include:
a statement of between 2 and 5 pages on (one of) the topics outlined above, a brief vitae.
The due date is August 15, 1996. We will respond by August 22.
Copyright 1997 Computing Research Lab.