We investigate how to augment such models with quantitative information, so that the model can tell us not only what a term can do, but (for instance) "in how many ways" (for nondeterministic calculi), "with what probability" (for probabilistic calculi), or "in how many steps of computation". We present a general construction of quantitative relational models and a generic soundness theorem for them, and indicate the links with known quantitative models such as probabilistic coherence spaces.

This is joint work with Jim Laird (Bath), Giulio Manzonetto (Paris 13) and Michele Pagani (Paris 13))

Far-134 (Video Conferencing Room)

Department of Computer Science