Recursive Distributed Representations：（递归的分布式表征）.pdfVIP

Recursive Distributed Representations Jordan B. Pollack Laboratory for AI Research Computer Information Science Department The Ohio State University 2036 Neil Avenue Columbus, OH 43210 (614) 292-4890 pollack@cis.ohio-state.edu ABSTRACT A long-standing difficulty for connectionist modeling has been how to represent variable-sized recursive data structures, such as trees and lists, in fixed-width patterns. This paper presents a connectionist architecture which automatically develops compact distributed representations for such compositional structures, as well as efficient access- ing mechanisms for them. Patterns which stand for the internal nodes of fixed-valence trees are devised through the recursive use of back-propagation on three-layer auto- associative encoder networks. The resulting representations are novel, in that they com- bine apparently immiscible aspects of features, pointers, and symbol structures. They form a bridge between the data structures necessary for high-level cognitive tasks and the associative, pattern recognition machinery provided by neural networks. 2 J. B. Pollack 1. Introduction One of the major stumbling blocks in the application of Connectionism to higher- level cognitive tasks, such as Natural Language Processing, has been the inadequacy of its representations. Both local and distributed representations have, thus far, been unsuit- able for capturing the dynamically-allocated variable-sized symbolic data-structures trad- itionally used in AI.