The goal of the project is to develop a powerful formalism for rule-based probabilistic programming, to study its meta-properties, to implement it, and to carry out some practical experiments.

We aim at introducing probabilistic strategies in ρLog calculus and to study semantics of such formalism. We also propose an extension of ρLog which will be capable to process probabilistic data. One of the main subgoals of the project is to study strong normalization property of probabilistic ρLog calculus. The obtained probabilistic calculus will be implemented as a prototype programming tool on top of PρLog, and some of its possible application areas will be explored.

Project members:

• Besik Dundua (Principal Investigator)