Educational Technology & Society 4 (2) 2001
ISSN 1436-4522

Human Cognition and Social Agent Technology

(Book review)

Reviewer: Hong Hong
Massey University, New Zealand

Book details:

Human Cognition and Social Agent Technology
Kerstin Dartenhahn (Ed.), University of Reading, United Kingdom
2000, John Benjamins Co. Address: PO Box 75577, 1070 AN Amsterdam, The Netherlands.
ISBN 9027251398 (448 pages)

The book " Human Cognition and Social Agent Technology " is a compilation of many researchers’ works.  It discusses combination aspects of human cognition with social agent technology. If you are interested in how to design social agent with the perspective of human cognition, this book will help you, to not only find the trend of the interdisciplinary areas research, but also the methodology for design and implementation of social agent. The important feature of this book is that it covers some technical issues of agent technology and social psychology on the human cognition, which is distinguished from books from computer science or from social psychology.  How the price for this feature is that the book covers quite wide interdisciplinary knowledge. 

Because of the wide coverage of interdisciplinary and each chapter direct from researchers’ deep research results, to understand the book may need good background knowledge in the interdisciplinary areas, even though the editor claimed the reader is not required to have a strong background in computer science, social science, or philosophy.

Each chapter in this book is a self-contained contribution focused on a particular aspect of socially intelligent agents.  Although the chapters can be ordered differently, the author grouped them based on their close relation.

The first three chapters deal with the story-telling systems. Chapter 1, Narrative Intelligence, discusses the narrative psychology that humans understand intentional behavior by organizing it into narrative and apply the narrative theory to design agents by providing narrative cues.  Chapter 2, Digital Augmentation of Keepsake Objects: A Place for Interaction of Memory, Story, and Self, describes the cognitive, social and psychological aspects of the Rosebud system.  It details the theoretical framework and its design, which base on the concept of tangible media that can merge the physical and the digital world and satisfy users’ social and cognitive needs.  Then the system implementation and user experience are described in detail.  Chapter 3, Children as Designers of Interactive Storytellers, presents the authoring environment - cSAGE (Storytelling Agent Generation Environment) that encourages story telling and story making as well as helps children to create their own interactive storytellers. The chapter discusses SAGE’s design and implementation, and analyses results of several pilot studies that examine children’s interaction with the SAGE’s storytellers, and their ability to build their own meaningful characters. 

Chapters 4 and 5 discuss about socially situated software agents, such as personal representatives and ‘conscious’ agents.  Chapter 4, Autonomous Synthetic Computer Characters as Personal Representatives, describes a framework that supports users to construct autonomous personal representative, and presents the development history and design of the framework, and the application of a prototype.  Chapter 5, “Conscious” and Conceptual Learning in a Socially Situated Agent, discusses Conscious Mattie that is a socially situated and “conscious” software agent. The authors describe architecture and functionality of Conscious Mattie in detail, and in particular focus on mechanisms that make the agent “conscious” and allow conceptual learning.  The authors aim to build a software agent that can act as a seminar coordinator and communicating by email.

Chapters 6 – 10 mainly discuss different architectures for agents from different perspective by using model of human social behavior and research results of human cognitive processes. Chapter 6, Emotionally Grounded Social Interaction, tries to answer how do agents that are embedded in large societies focus the enormous potential for interaction with other agents towards the most productive encounters. Therefore, the emotional states are grounded in a physiology that ensures appropriate feedback between behavior and affective states in the interactions of the agent with the world, so that an autonomous agent in terms of a synthetic physiology is modeled. The authors suggest to apply the synthetic physiology model to co-habited mixed reality that involves large heterogeneous societies, and in particular discuss the role of emotion with respect to the COMRIS (Co-Habited Mixed Reality Information Spaces) project.   Chapter 7, Architectural Requirements fro Human-Like Agents Both Natural and Artificial: What sorts of machines can love?, discusses emotion in Machines, explains why only an information processing machine can love, and presents the architecture for human-like agents.  The architectural layers comprise mechanisms for reactive, deliberative, self-management, and non-semantic control, these will improve the design artificial agents and understanding biological agents. Chapter 8, Connecting Reflection and Action: A Heterogeneous Multi-Agent Model, discusses how to combine two very different control approaches - explicit reasoning between alternatives and situated dynamic agent-environment interactions, to multi-agent environment.  The authors combined the advantages of both approaches in the project MACTA that studies a hybrid multi-agent architecture, and discuss the general issues and “consciousness” of such a multi-agent model.  Chapter 9, The Role of Evaluation in Cognition and Social Interaction, discusses about the role of evaluation in human cognition and social interaction.  The authors aim to identify evaluations and values as particular mental attitudes that are important for a goal-pursuing agent, and to explore their cognitive and social role within a knowledge representation and goal-pursuing system.  In the social domains, evaluation of oneself and others is important for the individual’s decision making. Chapter 10, The Ontogeny of the Social Self, presents a theory of the development of the social psyche of an agent. The authors discuss how an agent is integrated into a social system and becomes a social being, and how internal, psychological activities process inside an agent. The theory opens a novel paradigm for the investigation of psychodynamic phenomena.

Chapters 11 – 16 present several concrete research projects that address different aspects of design and implementations in socially situated interactive system. Chapter 11, Computational Embodiment: Agents as Constructed Complex Systems, intends to answer what are the properties that makes an agent social and how to construct flexible systems that will allow us to develop and test increasingly sophisticated concepts of intelligent agents in social environment.  The authors use the wrapping approach in the MUVE (multi-user virtual environments) to develop heterogeneous software and hardware environment for creating computationally reflective systems, and define necessary requirements for agents. Chapter 12, Are We Having Fun Yet? Using Social Agents in Social Domains, presents JULIA - a MUD social agent, which gives navigational and other assistance in a MUD (Multi-User Dungeon).  The authors discuss the domain of MUDs, the general capability of Julia, and her implementation. Chapter 13 The Emergence of Personality: How to Create Souls from Cells, discusses the design of software agents that people enjoy to interact with and treat as living, socially functioning beings.  Based on the author’s successful project that construct intelligent, lifelike, emotionally captivating and synthetic organisms by using a set of simple, biologically inspired building blocks, the author analyses how to combine the blocks to create organisms with personality, and how cells were made into souls. Chapter 14, Machine-Mediated Communication: Agents of Representation, discusses that intelligent agents, robots, and so on are vehicles of action (agents) and meaning (media) between human beings that imply them as virtual links between people.  The author emphasizes that in order to demonstrating their media functionality, creating believable personalities for agents should combine the art of impressing human senses and the art of programming machine behaviors.  Chapter 15, Agents as Artworks and Agent Design as Artistic Practice, from the perspective of the visual arts, argues that agent design aspects of the artistic methodology can complement conventional scientific methodologies.  The author analyses examples and his own work to emphasize the value of artistic methodologies to agent design. Chapter 16, Living with Socially Intelligent Agents: A Cognitive Technology View, discuss the optimization of the cognitive fit of humans and agents, and the impacts of agent technology on human social interaction and on human cognition, and the changes in who or what we are.  The authors analyses costs and rewards of agent technology for humans, advantages and disadvantages of humans adapting to interaction with social agents.

In summary, the book is good for interdisciplinary audiences who are interested in the socially intelligent agents area and may have some background knowledge in the related areas, to understand what are happening in this interdisciplinary area. However, audiences may need to pay attention to features of the book.  Because each chapter in the book is a self-contained contribution on a particular aspect of socially intelligent agents, and these materials come from many different researchers, although these chapters are put in a sequential order, concepts or ideas sometimes have big jumps, and let readers feel a little bit mess in some way.