AI research at CMU is closely integrated with other activities in the Computer Science Department, and to a major degree with ongoing research in the Psychology Department. Although there are over 50 faculty, staff and graduate students involved in various aspects of AI research, there is no administratively (or physically) separate AI laboratory. To underscore the interdisciplinary nature of our AI research, a significant fraction of the projects listed below are joint ventures between computer science and psychology.

The Robotics Project (the "Hand-Eye Project") evolved within the Stanford Artificial Intelligence Laboratory under the guidance of John McCarthy, Les Earnest, Jerry Feldman, and Tom Binford. Major efforts have been undertaken to isolate and solve fundamental problems in computer vision, manipulation, and autonomous vehicles. Generalized cones were introduced for modeling the geometry of 3-dimensional objects, and programs were constructed which learned structural descriptions of objects from laser-ranging data ("structured light"). Stereo vision and texture have been examined. Several generations of robot programming languages have resulted in AL, an intermediate-level language for commanding manipulation. A computer-controlled roving vehicle ("the cart") detected obstacles (using 9-eyed stereo) and planned paths to avoid them.

I am absolutely delighted to be able to join in this morning to offer my reflections on the occasion of the official beginning of the Robotics Institute. Beginnings are full of promise and potential. This one is no exception. What the Robotics Institute will become -- what effects it will have, both witting and unwitting -- are for the future to tell. What we all have now is a sense of adventure and anticipation.

At the recent AAAI conference at Stanford, it became apparent that many new AI research centers are being established around the country in industrial and governmental settings and in universities that have not paid much attention to AI in the past. At the same time, many of the established AI centers are in the process of converting from older facilities, primarily based on Decsystem-10 and Decsystem-20 machines, to a variety of newer options. At present, unfortunately, there is no simple answer to the question of what machines, operating systems, and languages a new or upgrading AI facility should use, and this situation has led to a great deal of confusion and anxiety on the part of those researchers and administrators who are faced with making this choice. In this article I will survey the major alternatives available at present and those that are clearly visible on the horizon, and I will try to indicate the advantages and disadvantages of each for AI work. This is mostly information that we have gathered at CMU in the course of planning for our own future computing needs, but the opinions expressed are my own.

Finally, abstraction can be extended to involve multiple complexity. In particular, one of the most costly behaviors levels, leading to a hierarchy of plans, each serving as a of the basic problem solving strategies is their inefficiency skeleton for the problem solving process at the next level in dealing with goal descriptions that include conjunctions. of detail. The search process at each level of detail can Because there is usually no good reason for the problem thus be reduced to a sequence of relatively simple solver to prefer to attack one conjunct before another, an subproblems of achieving the preconditions of the next incorrect ordering will often be chosen. This can lead to step in the skeleton plan from an initial state in which the an extensive search for a sequence of actions to try to previous step in the skeleton plan has just been achieved.

Massachusetts Institute Technology Press

A framework for inductive inference in logic is presented: a Model Inference Problem is defined, and it is shown that problems of machine learning and program synthesis from examples can be formulated naturally as model inference problems. A general, incremental inductive inference algorithm for solving model inference problems is developed. This algorithm is based on Popper's methodology of conjectures and refutations [II]. The algorithm can be shown to identify in the limit [3] any model in a family of complexity classes of models, is most powerful of its kind, and is flexible enough to have been successfully implemented for several concrete domains. The Model Inference System is a Prolog implementation of this algorithm, specialized to infer theories in Horn form.

Reading, Mass.: Addison-Wesley

This chapter from the Mycin book is a brief overview of van Melle's Ph.D. dissertation (Stanford, Computer Science), and is a shortened and edited version of a paper appearing in Pergamon-lnfotech state of the art report on machine intelligence, pp. 249-263. Maidenhead, Berkshire, U.K.: Infotech Ltd., 1981. Mycin Book (1984)

Specifically, it is concerned with control strategies governing the formation and refinement of partial hypotheses about the identity of an utterance that can guarantee the discovery of the best possible interpretation. We assume a system that contains the following components: a) A Lexical Retrieval component that can find the k best matching words in any region of an utterance subject to certain constraints and can be recalled to continue enumerating word matches in decreasing order of goodness (where possible constraints include anchoring the left or right end of the word to particular points in the utterance or to particular adjacent word matches).