Abstract
Lecture 1: Logic, Probability and Semantics: Introduction and Motivation
Conditional probability as the analogue of logical inference, probabilistic programs as distribution transformers, overview of developments in programming languages for machine learning, probabilistic $\lambda$-
Lecture 2: Background in Measure Theory and Integration
$\Sigma$-algebras, measures, extension theorems, Lebesgue integration, the Radon-Nikodym theorem, conditional expectation on continuous spaces, disintegration.
Lecture 3: The Lawvere-Giry Monad and Probabilistic Relations
The category of measure spaces, probabilistic mappings, the Lawvere-Giry monad, its Kleisli category, Markov kernels, Kozen's semantics for a language with while loops.
Lecture 4: Markov Processes, Bisimulation, and Logical Characterization
Probabilistic bisimulation, the logical characterization of bisimulation; this will include some descriptive set theory: analytic spaces, the unique structure theorem and smooth equivalence relations.
Lecture 5: Metrics for Markov Processes
Metrics between Markov processes, the Kantorovich metric aka the Wasserstein metric, lifting the metric from distributions to processes; metrics from a real-valued modal logic.
The first session of this mini course will take place on Monday, August 29th, 2016 4:30 pm – 5:30 pm; the second session of this mini course will take place on Tuesday, August 30th, 2016 3:00 pm – 4:00 pm; the third session of this mini course will take place on Wednesday, August 31st, 2016 3:00 pm – 4:00 pm; the fourth session of this mini course will take place on Thursday, September 1st, 2016 3:00 pm – 4:00 pm.