[edit]

Almost 10 years ago, in April 1985, Bill Gale and Daryl Pregibon organized the first Workshop on Artificial Intelligence and Statistics at Princeton, New Jersey. In his Preface to the edited volume Bill motivated the Workshop by noting that research in each field could benefit the other. The 15 papers presented at the first Workshop reflected a variety of interdisciplinary themes: task analysis of and expert advisory systems for statistical strategy, reasoning under uncertainty, supervised and unsupervised learning, heuristic search in pattern recognition, statistical tools for knowledge acquisition, and characterizing the behavior of `intelligent' systems through statistical analysis. Papers in subsequent workshops have further developed some of these themes and/or branched into other areas at the interface of AI and statistics.

It is difficult to know the influence of the AI and Statistics Workshops on research in the two communities. Suffice it to say that I have used these Workshops to guide, gauge, and `correct' my own work. I have heard several volunteer it as their favorite forum for discussion and debate, and in any case, it seems generally acknowledged as a workshop of very high quality. Finally, in contrast to 15 papers presented at the first Workshop in 1985, there are 62 papers, selected from a total of 111 submissions, being presented at the Fifth International Workshop on AI and Statistics in 1995. Apparently, many concur with Bill's motivation for the original Workshop.

There are many people that helped bring this Workshop to fruition. Hans Lenz, the Program Chair, was responsible for selecting the program committee, assigning papers, collecting reviews, and deciding the technical program. He also iterated with me in scheduling the presentations. Despite occasional logistical problems in the Nashville--Berlin link, it was a pleasure working with Hans. Hans and I thank those that provided submissions for review -- unfortunately, not all quality submissions could be accommodated. We thank the program committee for reviews and advise on paper selection. Daryl Pregibon handled finances, local arrangements, and provided input on scheduling -- if there was a General CoChair position, then the title would belong to Daryl. The Tutorial Chair, Prakash Shenoy, put together a wonderful tutorial program. Our tutorial presenters, David Aha, Trevor Hastie, Steffen Lauritzen, and Glenn Shafer deserve special thanks -- the time investment necessary for a good tutorial is very high, and we appreciate their efforts. My wife, Pat Fisher, and Jing Lin, Julio Ortega, and Doug Talbert of Vanderbilt University helped put together the proceedings. Robyn Landers of the University of Waterloo has been of great service by maintaining the AI and Statistics electronic mailing list, over which considerable Workshop information has been promulgated.

Finally, I greatly appreciate the sage advice of past organizers, particularly their input on program committee selection, but on a variety of other issues as well. They include Peter Cheeseman, Bill DuMouchel, David Hand, Wayne Oldford, and Daryl Pregibon. I am optimistic that this year's meeting will follow in the tradition of high quality that they helped hone.

Doug Fisher, General Chair