Argumentation-based Inference and Decision-Making – A Medical Perspective John Fox 1,2 , David Glasspool 3 , Dan Grecu 2 Sanjay Modgil 1,2 , Matt South 2,, Vivek Patkar 2 1, Oxford University ,2 CRUK and ASPIC development group, 3 Edinburgh University Abstract This paper discusses argumentation in inference and decision-making from the perspective of requirements for argumentation services in medicine. A body of work centred on applications of argumentation in biomedicine, such as risk assessment and treatment planning, has led to a comprehensive view of argumentation as a form of evidential reasoning. This in turn has stimulated the development of a general formalisation of argumentation for reasoning and decision-making, which has been used as the foundation for a number of tools for modelling and supporting medical decision-making and workflow management. Over a similar period ideas of ȃadversarialȄ argumentation and non-monotonic logic have become established, notably DungȂs seminal calculus of opposition. The diverse approaches to argumentation led to the EU-funded ASPIC project which aims to develop a theoretical consensus on argumentation and to translate this consensus into practical standards and tools. The second part of the paper presents some results of the ASPIC project focusing particularly on inference and decision- making. Introduction In recent years argumentation has developed from being an informal topic largely of interest to lawyers, philosophers etc. to a formal one that is finding a major place in the burgeoning fields of cognitive science and information technology. Historically, the first communities that were concerned with argumentation, were those interested in its relationship with reasoning by individuals (notably Toulmin and his followers) and those interested in natural dialogues and rhetoric (notably the Amsterdam school and the work of Walton and Krabbe). With the growing interest in argumentation in computer science recent work has become more formal, with both theoretical and practical aspects being explored in fields as diverse as mathematical logic, software engineering and AI.