Published November 12, 2025 | Version 1.0.13
Software Open

GenJAX: Probabilistic Programming with Vectorized Programmable Inference

Authors/Creators

  • 1. Massachusetts Institute of Technology
  • 2. MIT CHI FRO
  • 3. Yale University

Description

GenJAX is a probabilistic programming language (PPL) designed around programmable inference - automation that allows users to express and customize Bayesian inference algorithms. The system provides modeling language automation for constructing complex probability distributions and inference automation for constructing Monte Carlo samplers and variational inference methods. This repository includes the GenJAX implementation, extensive documentation, and case studies from the POPL 2026 paper 'Probabilistic Programming with Vectorized Programmable Inference'.

Files

femtomc/genjax-v1.0.13.zip

Files (967.4 kB)

Name Size Download all
md5:a355156cdb296a2075e698b27492b819
967.4 kB Preview Download

Additional details

Related works

Is supplement to
https://github.com/femtomc/genjax (URL)

Software

Repository URL
https://github.com/femtomc/genjax

References

  • POPL 2026: Probabilistic Programming with Vectorized Programmable Inference