Parallelizable Feynman-Kac Models for Universal Probabilistic Programming

This feed contains four research papers. 1) "Parallelizable Feynman-Kac Models for Universal Probabilistic Programming" establishes a formal, expectation-based semantics for universal probabilistic programs (PPGs), proves finite-trace approximation bounds, and introduces VPF, a vectorized particle filter consistent with the semantics—relevant to correctness and trust in probabilistic inference engines. 2) "Let Functions Speak" proposes F<:DR, adding domain/range projection types to give safe, lightweight polymorphism (avoiding reliance on unsafe any in practice), which strengthens static type/