A major strength of frame-based knowledge representation languages is their ability to provide the knowledge base designer with a concise and intuitively appealing means expression. The claim of intuitive appeal is based on the observation that the object -centered style of description provided by these languages often closely matches a designer's understanding of the domain being modeled and therefore lessens the burden of reformulation involved in developing a formal description. To be effective as a knowledge base development tool, a language needs to be supported by an implementation that facilitates creating, browsing, debugging, and editing the descriptions in the knowledge base. We have focused on providing such support in a SmallTalk (Ingalls, 1978) implementation of the KL-ONE knowledge representation language (Brachman, 1978), called KloneTalk, that has been in use by several projects for over a year at Xerox PARC. In this note, we describe those features of KloneTalk's displaybased interface that have made it an effective knowledge base development tool, including the use of constraints to automatically determine descriptions of newly created data base items.

Eurisko is an AI program that learns by discovery. We are applying Eurisko to the task of inventing new kinds of three- dimensional microelectronic devices that can then be fabricated using recently developed laser recrystallization techniques. Three experiments have been conducted, and some novel designs and design rules have emerged. The paradigm for Eurisko's exploration is a loop in which it generates a new device configuration, computes its I/O behavior, tries to "parse" this into a functionally it already knows about and can use, and then evaluates the results. In the first experiment, this loop took place at the level of charged carriers moving under the effects of electric fields through abutted regions of doped and undoped semiconductors. Many of the well-known primitive devices were synthesized quickly, such as the MOSFET, Junction Diode, and Bipolar Transistor. This was unsurprising, as they were short sentences in the descriptive language we had defined (a language with verbs like Abut and ApplyEField, and with nouns like nDoped Region and IntrinsicChannellRegion).

The acquisition of expert knowledge is fundamental to the certain of expert systems. The conventional approach to building expert systems assumes that the knowledge exists, and that it is feasible to find an expert who has the knowledge and can articulate it in collaboration with a knowledge engineer. This article considers the practice of knowledge engineering when these assumptions can not be strictly justified. It draws on our experiences in the design of VLSI design methods, and in the prototyping of an expert assistant for VLSI design. We suggest methods for expanding the practice of knowledge engineering when applied to fields that are fragmented and undergoing rapid evolution. We outline how the expanded practice can shape and accelerate the process of knowledge generation and refinement. Our examples also clarify some of the unarticulated present practice of knowledge engineering.

BBN's project in knowledge representation for natural language understanding is developing techniques for computer assistance to decision maker who is collecting information about and making choices in a complex situation. In particular, we are designing a system for natural language control of an intelligent graphics display. This system is intended for use in situation assessment and information management.

Two closely related aspects of artificial intelligence that have received comparatively little attention in the recent literature are research methodology, and the analysis of computational techniques that span multiple application areas. We believe both issues to be increasingly significant as Artificial Intelligence matures into a science and spins off major application efforts. It is imperative to analyze the repertoire of AI methods with respect to past experience, utility in new domains, extensibility, and functional equivalence with other techniques, if AI is to become more effective in building upon prior results rather than continually reinventing the proverbial wheel. Similarly, awareness of research methodology issues can help plan future research buy learning from past successes and failures. We view the study of research methodology to be similar to the analysis of operational AI techniques, but at a meta-level ; that is, research methodology analyzes the techniques and methods used by the researchers themselves, rather than their programs, to resolve issues of selecting interesting and tractable problems to investigate, and of deciding how to proceed with their investigations. A public articulation of methodological issues that typically remain implicit in the literature may provide some helpful orientation for new researchers and broaden the perspective of many AI practitioners.

Research by members of the Department of Computer Science at Rutgers, and by their collaborators, is organized within the Laboratory for Computer Science research(LCSR). AI and AI-related applications are the major area of research within LCSR, with about forty people-faculty, staff and graduate students-currently involved in various aspects of AI research.

Artificial intelligence is that part of computer science that concerns itself with the concepts and methods of symbolic inference and symbolic representation of knowledge. Its point of departure -- it's most fundamental concept -- is what Newell and Simon called (in their Turing Award Lecture) "the physical symbol system." But within the last fifteen years, it has concerned itself also with signals -- with the interpretation or understanding of signal data. AI researchers have discussed "signal-to symbol transformations," and their programs have shown how appropriate use of symbolic manipulations can be of great use in making signal processing more effective and efficient. Indeed, the programs for signal understanding have been fruitful, powerful, and among the most widely recognized of AI's achievements.

In this article we discuss a method for learning useful conditions on the application of operators during heuristic search. Since learning is not attempted until a complete solution path has been found for a problem, credit for correct moves and blame for incorrect moves is easily assigned. We review four learning systems that have incorporated similar techniques to learn in the domains of algebra, symbolic integration, and puzzle-solving. We conclude that the basic approach of learning from solution paths can be applied to any situation in which problems can be solved by sequential search. Finally, we examine some potential difficulties that may arise in more complex domains, and suggest some possible extensions for dealing with them.

Work on expert systems has received extensive attention recently, prompting growing interest in a range of environments. Much has been made of the basic concept and of the rule-based system approach typically used to construct the programs. Perhaps this is a good time then to review what we know, asses the current prospects, and suggest directions appropriate for the next steps of basic research. I'd like to do that today, and propose to do it by taking you on a journey of sorts, a metaphorical trip through the State of the Art of Expert Systems. We'll wander about the landscape, ranging from the familiar territory of the Land of Accepted Wisdom, to the vast unknowns at the Frontiers of Knowledge. I guarantee we'll all return safely, so come along....

Amidst the beautiful foliage in Jackson, N.H., the each session circulated a position paper to the group, Second KL-ONE Workshop was held over a five-day raising the questions he wanted to see addressed at the period this past October. Not only did "KloneTalk" (a version of KL-ONE implemented in we have a general conference session, wherein people SmaiiTa k at Xerox PARC -- this inchrded a videotaped could report on activities at their own institutions, discuss demonstration of the system's interface), prototypes in issues of general interest, etc., we also had a knowledge representation, translation of INTERLISP two-and-a-half day working research session. KL-ONE to FranzLisp, a calculus of Structural The technieai discussion part of the Workshop Descriptions, and the KL-ONE Classifier, not to mention several others. We also had the larger group break up preceded the general conference, so that we could report into smaller working groups to consider inference in on findings to the larger group of participants (forty-six KL-ONE, representing beliefs, some KL-ONE practice this year, from twenty-one institutions). Also, we planned to cover only a small extensive Proceedings of the Workshop.