POPL 2025
Sun 19 - Sat 25 January 2025 Denver, Colorado, United States
Sun 19 Jan 2025 09:43 - 09:58 at Peek-A-Boo - First session Chair(s): Matthijs Vákár

Laziness has long been recognized as useful for inference in probabilistic programs [Koller et al. 1997, Pfeffer 2009, Kiselyov and Shan 2009, Pfeffer et al. 2015], as it can drastically reduce the space of possible program executions, automatically marginalizing random choices that cannot affect the program’s result. However, inference engines for modern PPLs have largely moved away from lazy evaluation. This abstract revisits the question of lazy inference in light of modern advances in PPL semantics [Dash et al. 2023] and implementation [Holtzen et al. 2020]. Concretely, we propose a new variant of knowledge compilation—a state-of-the-art approach to inference for discrete probabilistic programs [Holtzen et al. 2020, Kimmig et al. 2011]—that exploits lazy evaluation to significantly improve performance, and we prove it correct using Dash et al. [2023]’s semantics for lazy, higher-order probabilistic programming. Early experiments show that lazy knowledge compilation can deliver significant performance gains, suggesting that the insights of early PPLs can be fruitfully combined with modern advances.

Sun 19 Jan

Displayed time zone: Mountain Time (US & Canada) change

09:00 - 10:30
First sessionLAFI at Peek-A-Boo
Chair(s): Matthijs Vákár Utrecht University
09:00
5m
Day opening
Welcome
LAFI
Matthijs Vákár Utrecht University, Atılım Güneş Baydin University of Oxford
09:06
20m
Industry talk
Industry Talk: Basis - Programming Languages as Core Technology for AI
LAFI
09:27
15m
Talk
Towards Symbolic Execution for Probability and Non-determinism
LAFI
Jack Czenszak Northeastern University, John Li Northeastern University, Steven Holtzen Northeastern University
File Attached
09:43
15m
Talk
Lazy Knowledge Compilation for Discrete PPLs
LAFI
Maddy Bowers Massachusetts Institute of Technology, Alexander K. Lew Massachusetts Institute of Technology, Joshua B. Tenenbaum Massachusetts Institute of Technology, Vikash K. Mansinghka Massachusetts Institute of Technology, Armando Solar-Lezama Massachusetts Institute of Technology
09:59
15m
Talk
Reasoning About Sampling Without Sampling: Atomic Machines for Contextual Equivalence in Probabilistic Programs
LAFI
Anthony D'Arienzo University of Illinois and Sandia National Laboratories, Jon Aytac Sandia National Laboratories
10:15
15m
Talk
Exact Inference for Nested Discrete Probabilistic Programs (Remote)
LAFI
File Attached