AN AUGMENTED STATE TRANSITION NETWORK ANALYSIS PROCEDURE Daniel G. Bobrow Bolt, Beranek and Newman, Inc. Cambridge, Massachusetts Bruce Eraser Language Research Foundation Cambridge, Massachusetts Summary A syntactic analysis procedure is described which obtains directly the deep structure information associated with an input sentence. The implementation utilizes a state transition network characterizing those linguistic facts representable in a context free form, and a number of techniques to code and derive additional linguistic information and to permit the compression of the network size, thereby allowing more efficient operation of the system. By recognizing identical constituent predictions stemming from two different analysis paths, the system determines the structure of this constituent only once. When two alternative paths through the state transition network converge to a single state at some point In the analysis, subsequent analyses are carried out only once despite the ...

Stanford AI Memo 83 or Computer Science Technical Note 130, Computer Science Department, Stanford University, Stanford, CA, 1969

Reading, Mass.: Addison-Wesley.

This term is meant to suggest not only the usual specifics of programming system and language but also such more elusive and subjective considerations as ease and level of interaction, "forgivefulness" of errors, human engineering, and system "Initiative." In normal usage, the word "environment" refers to the "aggregate of social and cultural conditions that influence the life of an individual." The programmer's enivronment influences, to a large extent determines, what sort of problems he can (and will want to) tackle, how far he can go, and how fast. If the environment is "cooperative" and "helpful" -- the anthropomorphism is deliberate -- then the programmer can be more ambitious and oroductive. If not, he will spend most of his time and energy "fighting" the system, which at times seems bent on frustrating his best efforts.

PLANNER: A LANGUAGE FOR PROVING THEOREMS IN ROBOTS Summary Carl Project MAC - Massachuse PLANNER is a language for proving theorems and manipulating models in a robot. The language is built out of a number of problem solving primitives together with a hierarchical control structure. Statements can be asserted and perhaps later withdrawn as the state of the world changes. Conclusions can be drawn from these various changes in state. Goals can be established and dismissed when they are satisfied. The deductive system of PLANNER is subordinate to the hierarchical control structure in order to make the language efficient. The use of a general purpose matching language makes the deductive system more powerful. Preface PLANNER is a language for proving theorems and manipulating models in a robot.

The problem discussed in this paper, namely that of finding a function to satisfy a given argument-value table, is by no means new to computing science, or to mathematics. Thus, for example, the problem of fitting a curve to a set of points is a part of numerical analysis. However, I am concerned with finding a function over a non-metric space, and so my work is closer to that of Feldman et al. (1969) in what they call, 'grammatical inference' or to the automaton-synthesizing programs described by Fogel, Owens and Walsh (1966).

Grammars or syntax specifications address themselves to the characterisation in symbolic terms of the structure of complex expressions. Two types of expression of empirical interest have been studied: sentences in English and other'natural' languages, and programs written in some high-level procedural language like ALGOL. Expressions in these languages consist of sets of elements (words and characters) coordinated with one another according to the sensorily manifest relationship'alongside', more commonly termed'followed by'.

A robot, in order to act intelligently, must be able to reason from facts which its sensors detect to conclusions which govern its actions. This reasoning process is so central to human intelligence that it seems immediately relevant to the problems of robot design to consider its properties, how it might be analysed and imitated.

This paper is a survey and discussion of research work relevant to the task of constructing some kind of reasoning robot. The emphasis is entirely on the organization of the reasoning processes, in particular planning, rather than on hardware. In practice the reasoning would most probably be carried out within a digital computer. My objective is to clarify the relationship between some superficially rather disparate approaches to this task, and simultaneously to indicate what seem to be the key problem areas. No new experimental results are presented, but the approach to the subject which I have adopted is a consequence of earlier experimentation with a simple computer simulation of a robot (Doran 1968a, 1969).

A minimal:robot,Icnown as Freddy, has been constructed with the aim of connecting a usable device online to the Department's lc L 4130, under the Multi-Pop time-sharing system, and discovering the snags. (See figure 1). Various technical problems arise when such a device runs free. It is much easier to anchor it and allow it to push its world about. Our present world is a three-foot diameter sandwich of hardboard and polystyrene which is light and rigid.