Idealization of intelligence as an embodied activity, involving an integration of cognition, perception and the body, places the tightest constraints on the design space for AI artifacts, forcing AI to deeply understand the design tradeoffs and tricks that biology has developed. I propose that a step in the design of such artifacts is to broaden the notion of cognitive state from the current linguistic-symbolic, Language-of-Thought framework to a multi-modal one, where perception and kinesthetic modalities participate in thinking. This is in contrast to the roles assigned to perception and motor activities as modules external to central cognition in the currently dominant theories in AI and Cognitive Science. I develop the outlines of this proposal, and describe the implementation of a bimodal version in which a diagrammatic representation component is added to the cognitive state.
PERCEPTUAL REPRESENTATION AND REASONING B. Chandrasekaran and N. Hari Narayanan Laboratory for Artificial Intelligence Research Ohio State University Columbus, Ohio 43210 Abstract A common view of reasoning in cognitive science is that it is a process that operates on abstract sentential representations. This view implies a separation of reasoning from sensory perception. Consequently, the study of perception has proceeded relatively independently of the study of various reasoning strategies that humans employ. In this paper we argue that there are many commonsense situations in which human reasoning is tightly coupled with perception, in particular with perceptually represented experiential knowledge. This type of reasoning is referred to as perceptual reasoning. This idea is based on a proposal about representations and supporting mechanisms that underlie visual perception and imagery. Perceptual reasoning is explained in terms of experientially acquired perceptual inference rules. Fmally, the implications of this stance are discussed.
In this paper we present and discuss a conceptual framework to design and construct emotion-based agents. We hypothesize that these agents exhibit a sensorimotor intelligence capable of providing them competence to cope with real, aggressive, and unpredictable environments. We sustain that such intelligence -- based on a structured pictoric knowledge representation -- can bridge the gap between purely reactive agents and complex, logic based, reasoning systems. It is not at all surprising that we only need to apply our intelligence to understand the workings of our rational reasoning mechanisms. What is not so rational is to apply only our reasoning mechanisms to understand the workings of our surprising intelligence.
To survive in a dynamic and rich environment. For instance, when faced with the image of a moving object, the cognitive processor provides elements to recognition (is it a lion or a rabbit?), whereas the perceptual processor delivers an assessment of the prevailing color, moving speed, dimension, and other relevant features found in the scene (is it a huge object with a particular color - a predator, or a little quick moving object - a prey?). These characteristics compose a "perceptual image" which serves two purposes: on the one hand, it allows a rough evaluation of the situation and the corresponding decision making. On the other, it helps the search which underlies the process of recognition: instead of comparing the "cognitive image" under processing with all the elements stored in memory, the search is bound to those objects sharing the same perceptual image. To reach this desideratum, "cognitive images" and "perceptual images" extracted from the same source stimulus should be associated and memorized in such a way that the latter indexes the former.