This chapter focuses on ACT system that embodies the extremely powerful thesis that a single set of learning processes underlies the whole gamut of human learning—from children learning their first language by hearing examples of adult speech to adults learning to program a computer by reading textbook instructions. The computer simulation is called ACT. The ACT theory describes its application to research on abstraction of schemas. In ACT, knowledge is divided into two categories: declarative and procedural. The declarative knowledge is represented in a propositional network similar to semantic network representations.

This work is an attempt to formally represent the knowledge required for the solution of a difficult retrograde chess problem (figure I). This solution Includes the extension of a formal deductive system to Include an observational facility. FOL [9], we have detailed a proof of the solution of the puzzle, Including proofs for almost all of the necessary associated lemmas [2], We shall highlight the various representational decisions made In the process of axiomatiiing retrograde chess, discussing both the necessity for these particular choices, and their Implications for designers of representations for other domains. This work is part of the search for epistemologically effective formalisms for artificial Intelligence.

This approach leads to an infinite sequence of algorithms I. Surface Models

Artificial Intelligence, 12 (1), 23-40.

I will remain Ion8 and deeply indebted to Herb Simon, H. T. Kung, AI Newell, Jay

Doctoral dissertation, Computer ScienceDept., Stanford University.

... we began to program a computer understanding system thatwould attempt to process input texts. An item crucial to our ability to accomplishthis task was what we called a script. A script is a frequently repeated causalchain of events that describes a standard situation. In understanding, when it ispossible to notice that one of these standard event chains has been initiated,then it is possible to understand predictively. That is, if we know we are in arestaurant then we can understand where an "order" fits with what we justheard, who might be ordering what from whom, what preconditions (menu,sitting down) might have preceded the "order", and what is likely to happennext. All this information comes from the restaurant script.Hayes, J.E., D. Michie, and L. I. Mikulich (Eds.), Machine Intelligence 9, Ellis Horwood.

Our purpose in studying natural language understanding in conjunction with problem solving is to bring together the constraints of what formal representation can actually be obtained with the question of what knowledge is required in order to solve a wide range of problems in a semantically rich domain. We believe that these issues cannot sensibly be tackled in isolation. In practical terms we have had the benefits of an increased awareness of common problems in both areas and a realisation that some of our techniques are applicable to both the control of inference and the control of parsing. Early work on solving mathematical problems stated in natural language was done by Bobrow (STUDENT - (i]) and Chamiak (CARPS - [5]). However the rudimentary parsing and simple semantic structures used by Bobrow and Charniak are inadequate for any but the easiest problems. Our intention has been to build on B/RG Chris This work was supported by SRC grant number 94493 and an SRC research studentship for Mellish.

Summary of Ph.D. dissertation, Computer Science Dept., Stanford University (1979)."TEIRESIAS is a program designed to provide assistance on the task of building knowledge-based systems. It facilitates the interactive transfer of knowledge from a human expert to the system, in a high level dialog conducted in a restricted subset of natural language. This paper explores an example of TEIRESIAS in operation and demonstrates how it guides the acquisition of new inference rules. The concept of meta-level knowledge is described and illustrations given of its utility in knowledge acquisition and its contribution to the more general issues of creating an intelligent program."Also in:Readings in Artificial Intelligence, ed. Webber, Bonnie Lynn and Nils J. Nilsson, Palo Alto, CA: Tioga Publishing Co., 1981.Orig. in IJCAI-77, vol.1, pp. 321 ff. Preprint in Stanford HPP Report #HPP-77-9.See also: Artificial Intelligence, 12[#2]:409-427. Readings in Artificial Intelligence, ed. Webber, Bonnie Lynn and Nils J. Nilsson, Palo Alto, CA: Tioga Publishing Co., 1981

Chess Life and Review (February):84-85.