UCL JILL DANDO INSTITUTE OF SECURITY AND CRIME SCIENCE

Jill Dando Institute of Crime Science

• Home Page
• About Us
• People
• Postgraduate Taught Programmes
• Short Courses
• Research
• PhD Training - UCL SECReT
• JDI Centre for the Forensic Sciences
• Events
• Contact Us

Causal Reasoning with Causal Models

The causal interpretation of Bayesian networks has been, and remains, controversial. I shall here present such an interpretation. I present a criterion of type causality which depends upon an indeterministic understanding of Bayesian networks as opposed to the deterministic views of Pearl and Hausman. Nancy Cartwright has influentially denounced "Bayes net methods" on the grounds that they fail under unfaithfulness, which has led to a variety of demonstrations of how widespread unfaithfulness is. I shall argue that this is a red-herring, ignoring the nature of causal reasoning with causal models. Jon Williamson has suggested that the failure of the Common Cause Principle makes causal discovery a doubtful enterprise; I argue rather that it is a problematic enterprise, which requires the invention and development of reliable model-building rules. I offer some plausible model-building rules, which appear to offer promise for using Bayesian networks to develop a unified account of type and token causality.

Speaker

My research is in artificial intelligence and the philosophy of science and the interrelation between the two. Of particular interest to me is the nature of cognition and learning. From the philosophy side this manifests itself in the study of confirmation theory and, in particular, Bayesian methods of learning and theory evaluation. The artificial intelligence side includes all varieties of machine learning projects, and particularly their analysis in Bayesian terms. I am also co-founder of Psyche: An Interdisciplinary Journal of Research on Consciousness.