Computational Imagery JANICE GLASGOW AND DIMITRI

After many years of neglect, the topic of mental imagery has recently emerged as an active area of research and debate in the cognitive science community. This article proposes a concept of computational imagery, which has patentlal applica-tions to problems whose solutions by humans involve the use of mental imagery. Computational imagery can be defined as the ability to represent, retrieve, and reason about spatial and visual information not explicitly stored in long-term memory. The article proposes a knowledge representation scheme for computational imagery that incorporates three representations: a long-term memory, descrip-tive representatlon and two working-memory representations, corresponding to the distinct visual and spatial components of mental imagery. The three repre-sentations, and a set of primitive functions, are specified using a formal theory of arrays and implemented in the array-based language Nial. Although results of studies in mental imagery provide initial motivation for the representations and functionality of the scheme, our ultimate concerns are expressive power, In-ferential adequacy, and efficiency. Numerous psychological studies have been carried out and several, often conflicting, models of mental imagery have been proposed. This article does not present another computational model for mental imagery, but instead treats imagery as a problem-solving paradigm in artificial intelligence (AI). We propose a concept of computational imagery, which has potential appli-cations to problems whose solutions by humans involve the use of mental imagery. As a basis for computational imagery, we define a knowledge representation scheme that brings to the foreground the most important visual and spatial properties of an image. Although psychological theories are used as a guide to these properties, we do not adhere to a strict cognitive model: Whenever possible, we attempt to overcome the limitations of th