A computer program capable of acting intelligently in the world must have a general representation of the world in terms of which its inputs are interpreted. Designing such a program requires commitments about what knowledge

Harvard University. The work toward attaining "artificial intelligence’ ’ is the center of considerable computer research, design, and application. The field is in its starting transient, characterized by many varied and independent efforts. Marvin Minsky has been requested to draw this work together into a coherent summary, supplement it with appropriate explanatory or theoretical noncomputer information, and introduce his assessment of the state of the art. This paper emphasizes the class of activities in which a general-purpose computer, complete with a library of basic programs, is further programmed to perform operations leading to ever higher-level information processing functions such as learning and problem solving. This informative article will be of real interest to both the general Proceedings reader and the computer specialist.-- The Guest Editor.

We argue that the analysis of agent#environmentinteractions should be extended to include the conventions and invariants maintained by agents throughout their activity. We refer to this thicker notion of environmentasalifeworld and present a partial set of formal tools for describing structures of lifeworlds and the ways in which they computationally simplify activity. As one speci#c example, we apply the tools to the analysis of the Toast system and showhowversions of the system with very di#erent control structures in fact implement a common control structure together with di#erent conventions for encoding task state in the positions or states of objects in the environment.

This thesis describes EM-ONE, an architecture for commonsense thinking capable of reflective reasoning about situations involving physical, social, and mental dimensions. EM-ONE uses as its knowledge base a library of commonsense narratives, each describing the physical, social, and mental activity that occurs during an interaction between several actors. EM-ONE reasons with these narratives by applying "mental critics, " procedures that debug problems that exist in the outside world or within EM-ONE itself. Mental critics draw upon commonsense narratives to suggest courses of action, methods for deliberating about the circumstances and consequences of those actions, and—when things go wrong—ways to reflect upon and debug the activity of previously invoked mental critics. Mental critics are arranged into six layers, the reactive, deliberative, reflective, self-reflective, self-conscious, and self-ideals layers. The selection of mental critics within these six layers is itself guided by a separate collection

In this paper the major components of every programming language are identified as: (1) the elementary program state-ment, (2) mechanisms for linking elementary statements together, (3) the means by which a program can obtain data inputs. Several alternative forms of each of these components are also described, compared and evaluated. Many examples, fre-quently from list processing languages, illustrate the forms described. The advantages, disadvantages and factors influencing the choice of a form of component for a language are dis-cussed, and the paper concludes with the suggestion that programming languages evolve toward one which will permit all the most convenient ways of structuring programs, organiz-ing systems and referencing data. 1.

This article examines Marvin Minsky's Society of Mind theory of human cognition. We describe some of the history behind the theory, review several of the specific mechanisms and representations that Minsky proposes, and consider related developments in Artificial Intelligence since the theory's publication.

A computer program, nanrod ADT (A Distinctly pirical Prover of eor), s been witt ich pr r groups. Its orgti is s11y hrtc corrting y of tiques of e h mtie pried Ist 1 thrm, is well is se is i viCle for testinS d lltS spl-e hristics. A deild dcrtion of e prr is sult by c of its perfore pridins y sits to e particur probl of a predure capable of prov a variety of tr m s dn. S ve fola for rez efforts al e lin, d cars [ reed rk prev[sly de.

A planning tool box is a set of software modules implementing different planning algorithms, techniques, and representation languages. A planning system designer can use these modules---and possibly additional own ones--- for building a whole family of generic or application planners. The paper explains the concept and outlines work on developing such a tool box. In particular, it identifies descriptions of planning domain characteristics as a topic in planning theory from which results are needed for building a tool box, yet are missing. Overture: The tool box idea, and the paper Writing effective programs for difficult problem classes, like planning, would normally require to narrow down the range of the program's applicability: The more generality is admitted, the more overhead results. In consequence, it makes sense to tailor a planning system such that it fits closely the intended application domain or class of applications. In research, concentration on a single planning paradi...

Recent advances in robotics leads us to consider, on the one hand, the notion of a kernel, a set of stable algorithms that drive developmental dynamics and, on the other hand, variable body envelopes that change over time. This division reverses the classic notion of a fixed body on which different software can be applied to consider a fixed software that can be applied to different kinds of embodiment. Thus, it becomes possible to study how a particular embodiment shapes developmental trajectories in specific ways. It also leads us to a novel view of the development of skills, from sensorimotor dexterity to abstract thought, based on the notion of a fluid body in continuous redefinition. 1

The objective of this thesis is to explore ways of describing agents in logical theories. The contribution is that the logical theories we build are a generalised form of logic programs. Like normal logic programs, these theories have an intuitive declarative reading and a procedural interpretation to guide the implementation of automatic devices and software. Both human beings and machines can reason about these logical theories. We employ the amalgamation of object and meta-logic programs to model notions such as beliefs, goals and agent's "mental" activities. But we also accommodate less usual notions such as reactivity, openness, activation of goals and preference encoding, that have proved to be essential in realistic models of agents. Four logic programming languages to program agent with those features are introduced. We use an event-based approach to model dynamic universes with changing properties, concurrency and synergistic effects. NOTE: This is a copy of the thesis with si...