ERIC Number: ED243711
Record Type: Non-Journal
Publication Date: 1984-Apr
Reference Count: N/A
Harmony Theory: Problem Solving, Parallel Cognitive Models, and Thermal Physics.
Smolensky, Paul; Riley, Mary S.
This document consists of three papers. The first, "A Parallel Model of (Sequential) Problem Solving," describes a parallel model designed to solve a class of relatively simple problems from elementary physics and discusses implications for models of problem-solving in general. It is shown that one of the most salient features of problem solving, sequentiality, can emerge naturally within a parallel model that has no explicit knowledge of how to sequence analysis. This model exploits a new type of parallel distributed processing that employs stochastic processors and is based on a formal mapping between parallel computation and thermal physics. The mathematical theory of this type of processing (harmony theory) is discussed in the second and third papers. Titles of these papers are, respectively, "The Mathematical Role of Self-Consistency in Parallel Computation" and "Harmony Theory: Thermal Parallel Models in a Computational Context." The latter paper discusses thermal models which employ stochastic processors and rely on a formal mapping between parallel computation and statistical physics. A special subclass of thermal models (harmony theory) is defined as the implementation-level description of a general mathematical frame work for studying cognition. This theory is presented at the computational, algorithmic, and implementation levels. (Author/JN)
Descriptors: Computation, Mathematical Models, Mathematics, Models, Physics, Problem Solving, Schemata (Cognition), Statistics, Theories
Requests for reprints should be sent to the authors at Institute for Cognitive Science, C-015; University of California, San Diego; La Jolla, CA 92093.
Publication Type: Reports - Research
Education Level: N/A
Sponsor: Office of Naval Research, Washington, DC. Personnel and Training Branch.
Authoring Institution: California Univ., La Jolla. Center for Human Information Processing.