Walter Reitman

Strategy control and directed forgetting

Signals to subjects to forget presignal items (F-items) can completely eliminate the proactive interference postsignal items (R-items) would normally suffer from the F-items. To determine what happens to F-items, a strategy-control procedure was developed to test memory for F-items without destroying forget-signal credibility. The subjects saw short lists of paired associates, some containing a signal to forget presignal pairs. After each list, a single stimulus probe tested memory for the appropriate response. Tested pairs were always R-pairs unless a prearranged cue informed the subjects that the stimulus was from an F-pair. Although F-pairs did not interfere with recall of R-pairs, there was substantial recall of and interference among F-pairs. The results support a set differentiation mechanism of directed forgetting.

Pattern recognition and pattern-directed inference in a program for playing go

Skilled human go play presumes the ability to recognize and make inferences from many different kinds of complex patterns. The go playing program described here deals with these various pattern recognition activities in terms of a small set of basic scanning and recognition mechanisms. All of the programs inference processes are defined with respect to these mechanisms.

Goals and plans in a program for playing Go

A program that plays Go provides a basis for analyzing possibilities for extending present AI conceptions of planning and goal structures to problems that are ill-structured, dynamic, multiperson, resource-bound, and highly interactive. The focus is on mechanisms for communicating information and control over time and among a number of interacting processes, in a flexible but coherent manner. Using the capabilities of current computer languages, it is possible to specify planning and goal structures, and appropriate conventions for them, that will accommodate the demands of these increasingly complex and sophisticated problem environments.

Perception and representation of spatial relations in a program for playing Go

Observation of human spatial perception suggests several characteristics that may serve as useful guidelines in the design of perception components for artificial intelligence systems. This paper describes one component designed in accordance with these principles, a web perception processor for a Go playing program. After a brief discussion of some basic strategic Go concepts, the paper describes the construction and modification of perceptual webs. Then, using examples from a real game situation, it shows how web structures, and changes in those structures, provide meaningfully organized representations of significant strategic properties in actual play. The paper concludes with some observations on the central significance of perceptual processing for Go programs and human Go players alike.

AUTONOTE: A personal information storage and retrieval system

This paper describes AUTONOTE, a personal storage and retrieval system designed for use by individuals working with large bodies of information. The user may enter a variety of textual materials and assign descriptors and phrases by which these materials may be retrieved. He has available mechanisms for deleting, replacing, linking, and hierarchically organizing text items. The system is operating on-line in a time-sharing environment, and can be utilized from a variety of terminals. Both use and implementation are discussed in detail, with special attention to utilization of AUTONOTE through an alphanumeric CRT display. Also mentioned are potential artificial intelligence extensions and the use of the system in a study of scientific problem solving.

PATTERN RECOGNITION AND PATTERN-DIRECTED INFERENCE IN A PROGRAM FOR PLAYING GO1

Skilled human Go play presumes the ability to recognize and make inferences from many different kinds of complex patterns. The Go playing program described here deals with these various pattern recognition activities in terms of a small set of basic scanning and recognition mechanisms. All of the programs inference processes are defined with respect to these mechanisms. The treatment of pattern similarity is conservative. Novel configurations are broken down into subpatterns, and subsequent inferences regarding such configurations are based upon the properties of the subpatterns and the relations among them. The basic functions carried out by the recognition and inference system include establishing appropriate multilevel pattern organizations; maintaining and updating pattern interrelations as the board situation changes; recognizing configurational identities and similarities; monitoring and filtering situational changes with respect to their implications for higher-level planning and problem solving functions; and providing predictively useful pattern information for planning. Although still very weak by human standards, the program is demonstrably stronger than previous Go programs. In particular, tests of its performance resulted in the first reported win by a program against a nonnovice human player.

Referential communication in AUTONOTE, a personal information retrieval system

This paper is concerned with the mechanisms of human intelligence and natural language communication, and with the design of an interactive computer program that incorporates and utilizes analogous mechanisms to improve man-machine communication. AUTONOTE2 is an improved personal information retrieval system. It includes, in addition to AUTONOTE (a presently running system), (1) mechanisms allowing the user to employ certain kinds of noun phrases to describe the items he wishes to store and retrieve; and (2) mechanisms enabling the system to maintain a map of what the user is referring to. Though extremely limited by comparison with the analogous human capabilities, these mechanisms add significantly to the descriptive power available to the user, and to the ease and efficiency of communication with the system. Furthermore, because AUTONOTE is a file oriented system, these additions add little to the present low cost of using the system. AUTONOTE2 also suggests new directions for software development for man-machine interaction, and it provides a practical testing ground for ideas about intelligence derived from observation of natural intelligence.