Neural-symbolic cognitive agents: architecture, theory and application

Abstract

In real-world applications, the eective integration of learn- ing and reasoning in a cognitive agent model is a dicult task. However, such integration may lead to a better under- standing, use and construction of more realistic multiagent models. Existing models are either oversimplied or require too much processing time, which is unsuitable for online learning and reasoning. In particular, higher-order concepts and cognitive abilities have many unknown temporal rela- tions with the data, making it impossible to represent such relationships by hand. In this paper, we develop and apply a Neural-Symbolic Cognitive Agent (NSCA) model for online learning and reasoning that seeks to eectively represent, learn and reason in complex real-world applications.