Probabilistic and Differentiable Programming

This project consists of research on building efficient probabilistic solvers and applying existing probabilistic solvers to different practical application domains, such as security and computer networks.

Systems

PSI solver

PSI is a probabilistic solver that takes as input a probabilistic program and automatically computes the exact probability density function (PDF) represented by that program

SPIRE

SPIRE is a synthesis system that automatically transforms a program into one that complies with a given privacy policy

DP-Finder

DP-Finder is a system that automatically derives lower bounds on the differential privacy enforced by algorithms

Publications

2021

DP-Sniper: Black-Box Discovery of Differential Privacy Violations using Classifiers

Benjamin Bichsel, Samuel Steffen, Ilija Bogunovic, Martin Vechev

IEEE S&P 2021

Slides
Talk
Paper
Code

2020

λPSI: Exact Inference for Higher-order Probabilistic Programs

Timon Gehr, Samuel Steffen, Martin Vechev

PLDI 2020

Paper

2018

Bayonet: Probabilistic Inference for Networks

Timon Gehr, Sasa Misailovic, Petar Tsankov, Laurent Vanbever, Pascal Wiesmann, Martin Vechev

PLDI 2018

Talk
Paper

Incremental Inference for Probabilistic Programs

Marco Cusumano-Towner, Benjamin Bichsel, Timon Gehr, Martin Vechev, Vikash K. Mansinghka

PLDI 2018

Talk
Paper

Fine-grained Semantics for Probabilistic Programs

Benjamin Bichsel, Timon Gehr, Martin Vechev

ESOP 2018

Slides
Paper

DP-Finder: Finding Differential Privacy Violations by Sampling and Optimization

Benjamin Bichsel, Timon Gehr, Dana Drachsler-Cohen, Petar Tsankov, Martin Vechev

ACM CCS 2018

Slides
Talk
Paper
Code

2017

Synthesis of Probabilistic Privacy Enforcement

Martin Kucera, Petar Tsankov, Timon Gehr, Marco Guarnieri, Martin Vechev

ACM CCS 2017

Talk
Paper

2016

PSI: Exact Symbolic Inference for Probabilistic Programs

Timon Gehr, Sasa Misailovic, Martin Vechev

CAV 2016

Paper