Some experts are likely to be fiercely critical because of omissions or errors. Others with tunnel vision are likely to miss the point. Rosalind Picard, with considerable courage, addresses a broad collection of themes, including the nature of motivation, emotions, and feeling; the detection of emotional and other affective states and processes; the nature of intelligence and the relationships between intelligence and emotions; the physiology of the brain and other aspects of human physiology relevant to affective states; requirements for effective human-computer interfaces in a wide range of situations; wearable devices with a range of sensing and communication functions; philosophical and ethical issues relating to computers of the future; and a brief encounter with theology. This is a book with a bold vision. Some readers will find it inspiring and mind stretching.
Sloman was one of the first in the AI community to write about the role of emotion in computing (Sloman and Croucher 1981), and I value his insight into theories of emotional and intelligent systems. Alas, Sloman's review dwells largely on some details related to unknown features of human emotion; hence, I don't think the review captures the flavor of the book. However, he does raise interesting points, as well as potential misunderstandings, both of which I am grateful for the opportunity to comment on. Sloman writes that I "welcome emotion detectors in a wide range of contexts and relationships, for example, teacher and pupil." This might sound innocuous, but its presumption of the existence of emotion detectors is not.
This is a set of notes relating to an invited talk at the cross-disciplinary workshop on Architectures for Modeling Emotion at the AAAI Spring Symposium at Stanford University in March 2004. The organisers of the workshop note that work on emotions "is often carried out in an ad hoc manner", and hope to remedy this by focusing on two themes (a) validation of emotion models and architectures, and (b) relevance of recent findings from affective neuroscience research. I shall focus mainly on (a), but in a manner which, I hope is relevant to (b), by addressing the need for conceptual clarification to remove, or at least reduce, the adhocery, both in modelling and in empirical research. In particular I try to show how a design-based approach can provide an improved conceptual framework and sharpen empirical questions relating to the study of mind and brain. From this standpoint it turns out that what are normally called emotions are a somewhat fuzzy subset of a larger class of states and processes that can arise out of interactions between different mechanisms in an architecture. What exactly the architecture is will determine both the larger class and the subset, since different architectures support different classes of states and processes. In order to develop the design-based approach we need a good ontology for characterising varieties of architectures and the states and processes that can occur in them. At present this too is often a matter of much ad-hocery. We propose steps toward a remedy.
We describe an agent architecture that integrates emotions, drives, and behaviors, and that focuses on modeling some of the aspects of emotions as fundamental components within the process of decision-making. We show how the mechanisms of primary emotions can be used as building blocks for the acquisition of emotional memories that serve as biasing mechanisms during the process of making decisions and selecting actions. The architecture has been implemented into an object-oriented framework that has been successfully used to develop and control several synthetic agents and which is currently being used as the control system for an emotional pet robot.