The Languages for Inference (LAFI) workshop aims to bring programming-language and machine-learning researchers together to advance all aspects of languages for inference: languages that offer built-in support for expressing probabilistic or differentiable models, and methods for inference and optimization over them, as programs, to ease reasoning, use, and reuse.
Topics include but are not limited to:
- Design of programming languages for probabilistic inference and/or differentiable programming
- Inference algorithms for probabilistic programming languages, including ones that incorporate automatic differentiation
- Automatic differentiation algorithms for differentiable programming languages
- Probabilistic generative modelling and inference
- Variational and differentiable modeling and inference
- Semantics (axiomatic, operational, denotational, games, etc) and types for probabilistic and/or differentiable programming
- Efficient and correct implementation
- Applications of probabilistic and/or differentiable programming
This year, LAFI is sponsored by BASIS (https://www.basis.ai/).
This program is tentative and subject to change.
Sun 11 JanDisplayed time zone: Brussels, Copenhagen, Madrid, Paris change
10:30 - 11:00 | |||
10:30 30mCoffee break | Break POPL Catering | ||
11:00 - 12:30 | |||
11:00 45mKeynote | Keynote LAFI | ||
11:45 10mTalk | Monte Carlo Analysis of Probabilistic Programs LAFI | ||
11:56 10mTalk | Verifying Sampling Algorithms via Distributional Invariants LAFI Daniel Zilken , Tobias Winkler RWTH Aachen University, Kevin Batz RWTH Aachen University, Joost-Pieter Katoen RWTH Aachen University File Attached | ||
12:07 10mTalk | Sequential Monte Carlo Program Synthesis with Refinement Proposals LAFI Maddy Bowers Massachusetts Institute of Technology, Mauricio Barba da Costa MIT, Xiaoyan Wang Massachusetts Institute of Technology, Joshua B. Tenenbaum Massachusetts Institute of Technology, Vikash Mansinghka Massachusetts Institute of Technology, Armando Solar-Lezama Massachusetts Institute of Technology, Alexander K. Lew Yale University | ||
12:18 10mTalk | A Word Sampler for Well-Typed Functions LAFI | ||
12:30 - 14:00 | |||
12:30 90mLunch | Lunch POPL Catering | ||
15:30 - 16:00 | |||
15:30 30mCoffee break | Break POPL Catering | ||
16:00 - 18:00 | |||
16:00 10mTalk | Multi-Agent Systems for Traceable Bayesian Workflow LAFI Xianda Sun University of Cambridge, Andrew D. Gordon Cogna and University of Edinburgh, Hong Ge University of Cambridge | ||
16:12 10mTalk | Grammar-Constrained LLM Generation for Reliable and Efficient Probabilistic Program Synthesis LAFI Madhav Kanda University of Illinois Urbana-Champaign, Shubham Ugare Meta, Sasa Misailovic University of Illinois at Urbana-Champaign | ||
16:24 10mTalk | Language-Model Probabilistic Programming for Improving Autoformalization via Cycle Consistency and Incremental Type-Checking LAFI Mauricio Barba da Costa MIT, Fabian Zaiser MIT, Katherine Collins MIT, Romir Patel MIT, Timothy O'Donnell , Alexander K. Lew Yale University, Joshua B. Tenenbaum Massachusetts Institute of Technology, Vikash K. Mansinghka Massachusetts Institute of Technology, Cameron Freer Massachusetts Institute of Technology | ||
16:35 80mPoster | Poster Session LAFI | ||
Accepted Papers
Call for Extended Abstracts
Call for Extended Abstracts — LAFI 2026
Submission deadline: October 30, 2025, AoE
Submission website: https://lafi26.hotcrp.com
We invite the submission of extended abstracts (2 pages + references + optional appendices) to the Languages for Inference (LAFI) workshop, colocated with POPL 2026.
LAFI aims to bring programming-language and machine-learning researchers together to advance all aspects of languages for inference. Topics include but are not limited to:
- Design of programming languages for probabilistic inference and/or differentiable programming
- Inference algorithms for probabilistic programming languages, including ones that incorporate automatic differentiation
- Automatic differentiation algorithms for differentiable programming languages
- Probabilistic generative modelling and inference
- Variational and differentiable modelling and inference
- Semantics (axiomatic, operational, denotational, games, etc) and types for probabilistic and/or differentiable programming
- Efficient and correct implementation
- Applications of probabilistic and/or differentiable programming
Dissemination of research. The workshop is informal, and our goal is to foster collaboration and establish a shared foundation for research on languages for inference. The proceedings will not be a formal or archival publication, and we expect to spend only a portion of the workshop day on traditional research talks.
Format. Uploads must be in PDF. Although no specific format is required, we suggest using an ACM template (either single- or double-column) in review mode, which adds line number annotations that reviewers can refer to when giving feedback.
Page limit: 2 pages of main content, excluding references and appendices. (Please note that reviewers are not required or expected to read appendices.)
Anonymity: submissions should be anonymized for peer review.
In line with the SIGPLAN Republication Policy, inclusion of extended abstracts in the program should not preclude later formal publication.
We strive to create an inclusive environment that does not demand traveling for presenters or participants.
| Sun 11 Jan 2026 Workshop Date |
| Thu 27 Nov 2025 Notification to Authors |
| Thu 30 Oct 2025 Submission Deadline |
Pedro H. Azevedo de Amorim
University of Oxford
United Kingdom
Jialu Bao
Northeastern University
Jürgen Cito
TU Wien
Austria
Poorva Garg
University of California, Los Angeles
India
Ilias Garnier
France
Alexander K. Lew
Yale University
United States
Hugo Paquet
Inria, École Normale Supérieure
France
William S. Moses
University of Illinois Urbana-Champaign
United States
Feras Saad
Carnegie Mellon University
United States
Fabian Zaiser
MIT
Yizhou Zhang
University of Waterloo
Canada
