Proceedings of the 16th conference on Computational linguistics - Volume 2

In my capacity as Chairman of the Program Committee of Coling 96, I wish to express my gratitude to all my colleagues who have contributed to the reading and selection of about 400 papers submitted for Coliug 96. The Program Committee functioned by subcommittees, and each one of 15 subcommittees was chaired by one of the Program Committee members. The chairperson of each subcommittee chose the members of the subcommittee, and assigned two members to each paper to be reviewed. More than 150 members were finally involved in the hard and sometimes painful task of selecting (and rejecting !) papers from such high quality papers. Computational linguistics is, as its very nature, an interdisciplinary field. It is becoming more and more so with rapid expansion and development of the field. When I was asked to be the Chairman of the Program Committee, I was, honestly, scared of the vast and diverse topics to be covered and was not at all sure whether papers in such diverse fields could be judged in an appropriate way. The very nature of Coling as a single, truely international conference of the field also scared me. It implies that papers are submitted from geographically very scattered areas and that the reviewing process involves reviewers scattered around the world. In fact, we received papers from 39 countries and had reviewers from 21 different countries. However, these fears were resolved, I believe, satisfactorily by generous devotion of the members of the Program Committee and subcommittees. Each of the 15 subcommittees played the role of Program Committee for a conference of a moderate size. The danger of independent subcommittees is that Coling may become a collection of independent smaller conferences, which in turn may not reflect the intricate mutual relationships among subfields in Computational Linguistics (think about a paper on "discourse model for machine translation" !). I believe that this danger was avoided by cross-reviewing the same papers, and by the co-operative attitude of the Program Committee members involved. In the Call for Papers, I abolished the distinction between topical papers and project notes. This is because I observed the distinction between these two categories had become somewhat blurred, and sometimes caused unnecessary confusions. However, quite a few papers were rejected with such comments as "the paper is very informative though it does not propose a new research idea","interesting for deinonstration", etc. Since Coling, the largest conference of our field, should play the role of a platform for exchanging first-hand information, I decided to have poster/demonstration forums for these papers. I was pleased that most of the authors of these papers agreed to present their papers in these forums, even though my decision was made after the first letters of acceptance were sent out. In the panels, we tried to cover topics which could not be covered by research papers but which are crucial for our field, such as social implications of the technology originating out from our research, the funding situation for our field, etc. For some reason, the nuinber of papers in fields such as language processing in speech systems, new applications of NLP, etc. are much less than I expected and hoped. Some of the panels address these issues and I hope that, these panels will stimulate new research in these areas. We adopted anonymous review for the first time in Coling. It seems, though this is not certain, that this had the effect of diversifying the research groups of the accepted papers. We also allowed e-mail submissions. Though this caused some problems and inconvenienced the authors of some papers, it worked fairly well.