ERIC Number: ED239868
Record Type: Non-Journal
Publication Date: 1983-Dec
Pages: 26
Abstractor: N/A
ISBN: N/A
ISSN: N/A
EISSN: N/A
Harmony Theory: A Mathematical Framework for Stochastic Parallel Processing.
Smolensky, Paul
This paper presents preliminary results of research founded on the hypothesis that in real environments there exist regularities that can be idealized as mathematical structures that are simple enough to be analyzed. The author considered three steps in analyzing the encoding of modularity of the environment. First, a general information processing task for a cognitive system, "completion," is described. Second, the entities of schemas for encoding the environmental modules and how they are used by the cognitive system to perform its task are discussed. Finally, a criterion for how the encoding of the environment into schemas should be done is presented. The paper is divided into sections on Completions and Schemas, Harmony and Computational Temperature, Cooling the System, Cognitive Crystallization, Schema Selection, Perceptual Grouping, Rules versus Instances, and Higher Level Analyses. An appendix presents the Formal Framework of Harmony Theory. A list of other reports and personnel is also included. (MNS)
Publication Type: Reports - Research
Education Level: N/A
Audience: Researchers
Language: English
Sponsor: Office of Naval Research, Arlington, VA. Personnel and Training Research Programs Office.
Authoring Institution: California Univ., San Diego, La Jolla. Inst. for Cognitive Science.
Grant or Contract Numbers: N/A