The DUAL Cognitive Architecture: A Hybrid Multi-Agent Approach

A hybrid (symbolic/connectionist) cognitive architecture, DUAL, is proposed. It is a multi-agent system which consist of a large number of non-cognitive, relatively simple agents, and which behaviour emerges from the behaviour of these simple agents and the interactions between them. The agents within this architecture have no internal knowledge base and goals. They are both computational devices and representational elements. They have internal (local) memory and hardwired processes that they can run. DUAL is hybrid at the micro level (i.e. it consists of hybrid agents) rather than at the macro level (i.e. it does not consist of separate symbolic and connectionist modules). From the symbolic perspective each agent represents a piece of world knowledge and performs some specific task while from the connectionist perspective it computes simply the activation level of the agent which reflects its relevance to the context. In this way symbolism and connectionism are considered as dualistic aspects of human cognition: the former representing the world knowledge and the latter its current relevance. The connectionist aspect of the architecture continuously "restructures" the knowledge base of the cognitive system represented by the symbolic aspect thus controlling the set of possible inferences at any moment. This makes the knowledge base dynamic and context-sensitive. The use of the DUAL cognitive architecture in modelling similarity judgements, analogical and deductive reasoning is described.