Inductive Logic & Statistics
That Hume's problem of induction remained unsolved two centuries later was a scandal for Philosophy, so said Bertrand Russell. But equally scandalous is Hume's analysis of causation. Here, in the Philosophy Department at Carnegie Mellon, we revel in both of these intellectual challenges.
Causal inference, and its connections with statistical theory, is the focus of the important, ongoing research program, led by Peter Spirtes, Clark Glymour, and Richard Scheines. Their book, Causation, Prediction, and Search (2nd edition) reports their original theory, which provides the application of directed statistical graphs to discovery and testing of causal models using observational and experimental data. The TETRAD program allows advanced implementations. These researchers vigorously pursue their theory in directions as diverse as automated (robotic) spectroscopy, gene mapping, and public health policy. Research in the foundations of statistical inference is another interdisciplinary theme that the Department advances. The team of Teddy Seidenfeld (in Philosophy), and Joseph Kadane and Mark Schervish (in Statistics) have collaborated for 20 years on foundational questions, including: How do you make statistical decisions with more than one decision maker? What is the theory of finitely additive probability as it relates to statistics? Are there degrees of incoherent statistics, so that some errors are better than others? Some of their work appears in Technical Reports for the Statistics Department. Their recent book, Rethinking the Foundations of Statistics, contains a selection of their papers in decision theory.
As a third example of how the Philosophy Department is engaged in spreading Hume's scandals, these same faculty are part of the Center for Automated Learning and Discovery in the School of Computer Science, CALD, where they help to administer the Center's graduate programs in datamining. A novel opportunity for Ph.D. students in Philosophy that grows out of this interdisciplinary collaboration is the possibility of obtaining a specialized SCS Masters degree in Datamining, through CALD, in lieu of the regular Masters degree in Philosophy. An example of this is found in the work of our current Ph.D. student, Tianjiao Chu who was the first recipient of this new MS degree in Computer Science. If only Hume could see us now!