No próximo dia 14 de Novembro, pelas 14h30 na sala FC6 S3 do DCC FCUP, Gabriele Vanoni irá dar uma palestra intitulada "Higher-Order Bayesian Networks".
A palestra é organizada pelo DCC-FCUP.
Short Bio
Gabriele Vanoni defended his PhD thesis (supervised by Ugo Dal Lago) at Università di Bologna in June 2022. In his dissertation, he has solved a longstanding open problem about LOGSPACE computation in functional programming languages. For this result, he has been awarded the Beth Dissertation Award, the Ackerman Outstanding Dissertation Award, and the best Italian PhD thesis in theoretical computer science award. More in general, his research lies in the field of logical foundations of (functional) programming languages. After having spent one year at Inria in Sophia Antipolis as a postdoc, he is now a postdoctoral researcher at IRIF, in Paris.
Title
Higher-Order Bayesian Networks
Abstract
Bayesian networks (BNs) are graphical first-order probabilistic models that allow for a compact representation of large probability distributions, and efficient inference, both exact and approximate. We introduce a higher-order programming language, in the idealized form of a lambda-calculus, which we prove sound and complete w.r.t. BNs: each BN can be encoded as a term, and conversely each (possibly higher-order and recursive) program of ground type compiles into a BN. Moreover, we provide a compositional and cost-aware semantics which is based on factors, the standard mathematical tool used in Bayesian inference. Our results rely on advanced techniques rooted into linear logic, intersection types, rewriting theory, and Girard's geometry of interaction, which are here combined in a novel way. This is joint work with Claudia Faggian and Daniele Pautasso.