A rgumentation-which used to be an informal topic, of interest primarily to lawyers and philosophers-has become in recent years an area with a strong formal foundation, with numerous applications in cognitive science and information technology. The first communities concerned with argumentation were those interested in its relationship to reasoning by individuals (notably Stephen Toulmin and his followers) and those interested in natural dialogue and rhetoric (notably the Amsterdam school and Douglas Walton and Erik Krabbe's work). With the growing interest in argumentation in computer science, researchers have started to explore both theoretical and practical aspects in fields as diverse as mathematical logic, software engineering, and AI.From a theoretical perspective, argumentation offers a novel way of understanding and formalizing human cognition. Logicians and computer scientists are also using argumentation systems for expanding their investigations into classical forms of reasoning such as deduction and reasoning under uncertainty, and into nonclassical forms of reasoning such as nonmonotonic, modal, and epistemic logics. Another promising application area for argumentation theory is distributed AI and multiagent systems, where the dialectical nature of argumentation helps structure and reason about the exchange of knowledge and information related to claims. From a practical perspective, argumentation provides a versatile computational model for developing advanced services for decision support and for computerhuman dialogues in which explanation and rationale play a central role.Argumentation is a potentially important paradigm for developing commercial and public services that are flexible and easily understood by human users. In the first part of this article, we present our work on developing argumentation-based services for biomedical applications, with a particular focus on argument-based inference, decision making and planning. The second part of the article concerns the theoretical and practical lessons learned from this initial work and its follow-on as part of the Argumentation Services Platform with The Argumentation Services Platform with Integrated Components (ASPIC) project aims to provide advanced argumentation-based computational capabilities.

IntroductionDesign represents one of the most complex problem solving domains addressed by Artificial Intelligence. Despite the progress made in the last decade to advance the use of AI techniques in design, existing systems have difficulty coping with the diversity and quantity of knowledge required, as well as with the variety of reasoning involved.In general:• A design problem requires knowledge from various domains, and uses a broad range of representations.• Design problem solving is based on the ability to carry out many specialized tasks, such as analysis, abstraction, evaluation and explanation, each involving different reasoning abilities.Portions of some design domains have been analysed and formalized, providing solid support for the search for solutions. However, much of designing still relies on good knowledge and heuristics. Maintaining the quality of designs and the efficiency with which they are produced requires continual evaluation and improvement of design knowledge and methods, including heuristics.For designers, such improvements have been based on recording and learning from notable events and attributes that have occurred during the development of designs. Learning from designs, and learning during designing, is as old as design activity itself.Adding adaptation to a design system is clearly desirable. Even though learning does not always reach the optimal solution, experience should eventually bring noticeable and worthwhile improvements over initial designs and design processes. These are measured in terms of higher quality, shorter design times, and lower costs.There has been increasing acknowledgment that computational design systems can and should include the ability to learn, and there is an increasing amount of research on Machine Learning in Design (as demonstrated by the papers in this special issue).

No abstract

Design Expert Systems can be built using many small, cooperating, limited function expert systems called Single Function Agents (SiFAs). Using this approach we will be able to investigate and discover primitive problem-solving and interaction patterns, specific for multiagent design systems, and should gain a deeper understanding of the types of knowledge involved. This paper presents some categories of conflicts that have been studied using the SiFA approach, and makes a brief presentation of the SINE implementation of SiFAs.

No abstract