Caesar: A Verifier for Probabilistic Programs - Dafny'24
Offered By: ACM SIGPLAN via YouTube
Course Description
Overview
Explore a groundbreaking verifier for probabilistic programs in this 18-minute conference talk from ACM SIGPLAN. Delve into Caesar, a novel deductive verification infrastructure first introduced at OOPSLA '2023. Discover the basics of this infrastructure and its implementation, drawing parallels to Boogie as a probabilistic counterpart. Examine the key components, including an intermediate verification language (IVL) and a real-valued logic, designed to handle quantitative properties such as expected outcomes, run-times, and termination probabilities. Learn about the paradigm shift from Boolean to real-valued domains in verification, and understand how the IVL incorporates quantitative generalizations of standard verification constructs. Gain insights into the weakest-precondition-style semantics for generating verification conditions and the SMT-based implementation that enables automatic verification of various benchmarks. Recognize the significance of this infrastructure as the first of its kind for expectation-based reasoning about probabilistic programs.
Syllabus
[Dafny'24] Caesar: A Verifier for Probabilistic Programs
Taught by
ACM SIGPLAN
Related Courses
Statistical RethinkingMax Planck Institute for Evolutionary Anthropology via YouTube Introducción a las bases del lenguaje R, con RStudio
Udemy Bayesian Networks 1 - Inference - Stanford CS221: AI
Stanford University via YouTube The Hamiltonian Monte Carlo Revolution Is Open Source - Probabilistic Programming with PyMC3
Open Data Science via YouTube Computational Models of Cognition - Reverse-Engineering Common Sense in the Human Mind and Brain Pt 1
MITCBMM via YouTube