Procedural Knowledge Processing Based on Area Representation Using a Neural Network
Seiya Fujinaga and Masafumi Hagiwara
Department of Electrical Engineering, Faculty of Science and Technology, Keio University 3-141 Hiyoshi, Kohoku-ku, Yokohama 223-8522 Japan
In this paper, a neural network that treats procedural knowledge based on area representation is proposed. The main theme of this paper is to propose a novel neural network that processes procedural knowledge. The network employs formerly proposed ideas such as “area representation” and “improved Hebbian learning.” Area representation expresses information by a group of neurons. Since it is considered as a combination of localized and distributed representation, it has many advantages such as robustness, high efficiency for information representation, and potential ability to treat similarity of data. The proposed network based on area representation is constructed to store and recall procedural knowledge. We performed various kinds of computer simulations to examine the validity and effectiveness of the proposed network.
This article is published under a Creative Commons Attribution-NoDerivatives 4.0 International License.