Por Melissa Antonelli (University of Bologna).

Interactions between logic and theoretical computer science have been deeply investigated in the realm of deterministic computation but, strikingly, when switching to the probabilistic framework, the literature does not offer much. The main goal of our study is precisely to start bridging this gap by developing inherently quantitative logics and investigating their relations with specific aspects of randomized models. This talk is conceived as an overview of our work (in particular, of ongoing research) and is bipartite. In the first part, I will introduce our quantitative counting propositional logics, which are basically obtained by endowing standard propositional systems with counting quantifiers, expressing the probability of the (argument) formula. I will show that the classical fragment, CPL, is strongly linked to complexity theory, as characterizing the full counting hierarchy. On the other hand, in the context of programming language theory, the intuitionistic counting logic, iCPL_0, provides the first probabilistic correspondence in the style of Curry and Howard In the second part of the talk, I will introduce the more expressive logic MQPA, the language of which is defined by extending that of PA with measure-quantifiers. This logic is strongly connected to probabilistic computation, as shown by some results, such as “randomized arithmetization”. Starting from this, we are also developing randomized bounded theories, in order to logically characterize probabilistic complexity classes, following the path delineated by Buss’ and Ferreira’s bounded arithmetics. (Part of this work, which is still in progress, has been conducted during my visiting period at NOVA University.)

Transmissão via Zoom.