MIT's Probabilistic Computing Project

News

  • November 2025: Fast Controlled Generation from Language Models with Adaptive Weighted Rejection Sampling, co-authored by ProbComp members, was selected as one of four “Outstanding Papers” at this year’s Conference on Language Modeling (COLM). 
  • September 2025: Congratulations to Dr. Joao Loula for completing his PhD program.
  • July 2025: Our paper Understanding epistemic language with a language-augmented bayesian theory of mind has been accepted for publicatIon at TACL 2025.
  • July 2025: Our papers Tracking Uncertainty During Uncertain Tracking and Seeing through Occlusion: Uncertainty-aware Joint Physical Tracking and Prediction have been accepted for publciation at CogSci 2025.
  • June 2025: Our paper Stochastic Lazy Knowledge Compilation for Inference in Discrete Probabilistic Programs has been accepted for publication at PLDI 2025
  • May 2025: Congratulations to Drs. Alex Lew, Nishad Gothoskar, and Tan Zhi Xuan for completing their PhD programs.
  • April 2025: Our research paper Syntactic and Semantic Control of Large Language Models via Sequential Monte Carlo has been accepted for publication and an oral presentation at ICLR (meaning it’s rated among the top ~2% of papers this year).
  • October 2024: Our paper Building machines that learn and think with people has been published in Nature human behavior
  • June 2024: Two papers from the lab, Probabilistic Programming with Programmable Variational Inference and GenSQL: A Probabilistic Programming System for Querying Generative Models of Database Tables have been presented at PLDI in Copenhagen.
  • May 2023: Congratulations to Dr. Feras Saad, for winning the George M. Sprowls PhD Thesis Award in Artificial Intelligence and Decision-Making for his thesis Scalable Structure Learning, Inference, and Analysis with Probabilistic Programs.
  • January 2023: Our research paper SMCP3: SMC with Probabilistic Program Proposals has been accepted for publication at AISTATS 2023.
  • January 2023: Our research paper ADEV: Sound Automatic Differentiation of Expected Values of Probabilistic Programs has been presented at POPL 2023 and covered by MIT News.
  • November 2022: Our research paper Visual object detection biases escape trajectories following acoustic startle in larval zebrafish has been accepted for publication at current biology and covered by MIT News.
  • April 2022: Our research paper [Estimators of Entropy and Information via Inference in Probabilistic Models][7] has been presented at AISTATS 2022 and covered by [MIT News][8].
  • October 2021: Our research on [3DP3: 3D Scene Perception via Probabilistic Programming][9] has been accepted for publication at NeurIPS 2021 and covered by [MIT News][10].
  • August 2021: Our research on [probabilistic programming with fast exact symbolic inference][11] was covered by [MIT News][12]
  • May 2021: Our research on [Bayesian Data Cleaning at Scale with Domain-Specific Probabilistic Programming][13] was covered by [MIT News][14]
  • May 2021: We are happy to share that our paper [Hierarchical Infinite Relational Model][15] has been accepted to for publication and oral (long-form) presentation at UAI 2021.
  • March 2021: We are pleased to announce that our paper [SPPL: Probabilistic Programming with Fast Exact Symbolic Inference][16] has been accepted to PLDI 2021.
  • February 2021: We are happy to announce that our paper [PClean: Bayesian Data Cleaning at Scale with Domain-Specific Probabilistic Programming][13] has been accepted for publication and oral presentation at AISTATS 2021.
  • We are pleased to share that Vikash Mansinghka is an invited keynote speaker at [LAFI 2021.][17]
  • January 2021: We are happy to announce that our paper [Transforming Worlds: Automated Involutive MCMC for Open-Universe Probabilistic Models][18] has been accepted at the 3rd Symposium on Advances in Approximate Bayesian Inference (AABI).
  • December 2020: Our research on [Online Bayesian goal inference for boundedly-rational planning agents][19] has been covered by [MIT News][20]
  • December 2020: We are pleased to share that we had two papers accepted at LAFI 2021:
    • [On the Automatic Derivation of Importance Samplers from Pairs of Probabilistic Programs.][21] Alexander K. Lew, Ben Sherman, Marco Cusumano-Towner, Michael Carbin, Vikash Mansinghka
    • [Genify.jl: Transforming Julia into Gen to enable programmable inference.][22] Tan Zhi-Xuan, McCoy R. Becker, Vikash Mansinghka
  • September 2020: We are happy to announce that our paper [Online Bayesian goal inference for boundedly-rational planning agents][19] has been accepted for publication at NeurIPS 2020.
  • June 2020: We are pleased to share that our paper [Causal inference using Gaussian processes with structured latent confounders][23] has been accepted for publication at ICML 2020.
  • May 2020: Our research on [The Fast Loaded Dice Roller][24] was covered by [MIT News][25], [Intel][26], and [Quanta][27].
  • January 2020: We are happy to announce that our paper [The Fast Loaded Dice Roller: A Near-Optimal Exact Sampler for Discrete Probability Distributions][24] has been accepted for publication at AISTATS 2020.
  • November 2019: We are pleased to announce that two papers have been accepted to [POPL 2020][28]:
  • [A Type System and Semantics for Sound Programmable Inference in Probabilistic Languages][29]. Alexander K. Lew, Marco F. Cusumano Towner, Benjamin Sherman, Michael Carbin, Vikash K. Mansinghka.
  • [Optimal Approximate Sampling from Discrete Probability Distributions][30]. Feras A. Saad, Cameron E. Freer, Martin C. Rinard, Vikash K. Mansinghka.
  • June 2019: Our research on Gen was covered on [MIT News][31] and further covered by [VentureBeat][32], [ZDNet][33], and featured on [Hacker News][34].
  • April 2019: We are happy to announce that our paper [Gen: A General-Purpose Probabilistic Programming System with Programmable Inference][35] was accepted to PLDI 2019.
  • January 2019: Our research on [Bayesian Synthesis of Probabilistic Programs for Automatic Data Modeling][36], presented at POPL 2019, was featured on [MIT News][37] and [ZDNet][38].
  • January 2019: We are happy to announce that our paper [A Family of Exact, Distribution-Free Goodness-of-Fit Tests for High-Dimensional Discrete Distributions][39] was accepted to AISTATS 2019.
  • November 2018: We are pleased to announce that our paper [Bayesian Synthesis of Probabilistic Programs for Automatic Data Modeling][36] was accepted to POPL 2019. See [here][40] for a video of the talk presented at the conference.
  • October 2018: We hosted the [Inaugural International Conference on Probabilistic Programming (PROBPROG 2018)][41] organized by Vikash Mansinghka, Jan-Willem van de Meent, and Avi Pfeffer. Lab members Marco Cusumano-Towner, Feras Saad, and Ulrich Schaechtle presented research talks and Vikash Mansinghka hosted a tutorial session. Videos of talks are available [here][42].
  • August 2018: We gave an invited talk at the Joint Statistical Meetings (JSM) 2018 on [Probabilistic Programming with Non-Parametric Bayesian Model Discovery in BayesDB][43].
  • June 2018: [Tableau acquired Empirical][44], the second startup based on components of the MIT Probabilistic Computing Stack, into their ‘automated intelligence’ products group, to provide automated data insights.
  • June 2018: The Defense Advanced Research Projects Agency (DARPA) selected Ulrich Schaechtle, post-doctoral researcher, as a DARPA Riser for 2018. Dr. Schaechtle will attend and present on machine-assisted causal modeling via probabilistic programming at [DARPA’s 60th Anniversary Symposium (D60)][45] in September at the Gaylord National Convention Center in National Harbor, Maryland.
  • May 2018: We gave an invited talk at the [AI for Good Global Summit][46] held by the UN International Telecommunications Union
  • April 2018: We are presenting lab member Dr. Ulrich Schechtle’s work on AI-assisted data search for collaborative science, using BayesDB, at the upcoming DARPA SD2 Working Meeting in San Jose, CA. This ongoing project is building on our pre-print [Probabilistic search for structured data via probabilistic programming and nonparametric Bayes][47].
  • April 2018: Our workshop paper titled *Gen: a probabilistic programming platform with fast custom inference via code generation* was accepted to [MAPL 2018 (Machine Learning and Programming Langauges Workshop)][48], affiliated with <a HREF=”https://conf.researchr.org/home/pldi-2018″>PLDI 2018</a>. Contact PhD student <a HREF=”mailto:marcoct@mit.edu”>Marco-Cusumano-Towner</a> for more information on Gen.
  • April 2018: We gave an invited keynote on AI-assisted data science with BayesDB at Deloitte’s annual Data Science Summit in Dallas, Texas, to 200+ data scientists and business leaders from across the organization.
  • March 2018: We gave an invited keynote on AI-assisted data science at the Alan Turing Institute’s Workshop on Artificial Intelligence for Data Analytics. A key theme we discussed was how probabilistic programming can help the open science and data science communities develop automated tools for searching data, screening data for quality errors, and helping to design more efficient experiments.
  • March 2018: We had two papers accepted to PLDI 2018: [Probabilistic Programming with Programmable Inference][49] and [Incremental Inference for Probabilistic Programming][50].
  • March 2018: We gave an invited presentation on BayesDB for AI-assisted data search at <a HREF=”https://ire.org/conferences/nicar18/”>the Investigative Reporters and Editors’ 2018 NICAR conference on computer assisted reporting</a>. We are excited to be collaborating with leading newsrooms on data search and data screening; for more information, please contact lab member <a HREF=”mailto:srsmith@mit.edu”>Sara Rendtorff-Smith</a>.
  • January 2018: Our paper [A Bayesian Nonparametric Method for Clustering, Imputation, and Forecasting in Multivariate Time Series][51] was accepted to [AISTATS 2018][52].
  • January 2018: We presented [Using probabilistic programs as proposals][53] at the [Probabilistic Programming Semantics (PPS) Workshop 2018][54], co-located with POPL.
  • December 2017: We are pleased to have presented a [tutorial on probabilistic programming][56] to over 4000 participants at the NIPS 2017 conference.
  • September 2017: Our paper [AIDE: An algorithm for measuring the accuracy of probabilistic inference algorithms][57] was accepted to NIPS 2017.
  • August 2017: Invited talk at the [Sequential Monte Carlo Workshop 2017][58] in Uppsala, Sweden.
  • July 2017: We are pleased to be teaching at the [Probabilistic Programming for Advanced Machine Learning][59] summer school hosted by DARPA in Arlington, Virginia. The focus will be on AI-assisted data analysis with BayesDB — learning how to answer data analysis questions in minutes that normally take hours or days for experienced computational statisticians.
  • June 2017: We presented tutorials on [Probabilistic Programming][60] and [AI for Structured Business Data][61] to an audience of over 200 participants at the O’Reilly Artificial Intelligence Conference, New York City.
  • June 2017: Vikash presented at the Technology Strategy and Leadership (TSLRP) meeting at Intel, including Intel’s CEO, management committee, and [fellows][62]. We briefed the company on new opportunities in probabilistic AI, including our work on building analogs of TenserFlow and the GPU for probabilistic programing.
  • June 2017: Invited talk on [AI Assisted Data Analysis for Humanitarian Causes][63] at Effective Altruism Global in Boston, MA.
  • May 2017: Two new preprints, [Time Series Structure Discovery via Probabilistic Program Synthesis][64] and [AIDE: An algorithm for measuring the accuracy of probabilistic inference algorithms][57], are now available on the arXiv.
  • May 2017: We taught an AI-assisted data science workshop at the Boston Children’s hospital, using BayesDB to analyze data relevant for understanding the effect of childhood adversity on the developing brain; in collaboration with the [Laboratories for Cognitive Neuroscience][65].
  • May 2017: Invited talk at [AI Advance][66], hosted by the [Ethics and Governance of Artificial Intelligence Initiative][67] at the Berkman Klein Center for Internet and Society, Harvard University.
  • April 2017: We presented [Detecting dependencies in sparse, multivariate databases][68] at the AISTATS conference in Fort Lauderdale, Florida.
  • March 2017: Check out our new pre-prints on [Probabilistic search for structured data via probabilistic programming and nonparametric Bayes][47] and [Probabilistic programs for inferring the goals of autonomous agents][69].
  • April 2017: Invited talk on AI-assisted data science at [DALI 2017][70] in Tenerife, Canary Islands.
  • January 2017: Participated in the DARPA Hackathon in Arlington, Virginia.
  • We used automated model discovery in BayesDB and Gaussian process modeling in Venture for a collection of challenging data science tasks: imputing missing data, modeling the effect of intervention variables on outcome variables, modeling missingness processes, and statistical model criticism. Our solutions to these problems, developed over the course of two days, were competitive with the baseline solutions developed by statistics specialists over the span of several weeks.
  • January 2017: Presented [Encapsulating models and approximate inference programs in probabilistic modules][71] at the POPL Workshop on Probabilistic Programming Semantics in Paris, France.
  • December 2016: We have presented the following collection of conference and workshop papers at the NIPS conference in Barcelona, Spain:
  • [A Probabilistic Programming Approach To Probabilistic Data Analysis][72], main conference;
  • [Gaussian process structure learning via probabilistic inverse compilation][73], Workshop on Interpretable Machine Learning in Complex Systems;
  • [Measuring the non-asymptotic convergence of sequential Monte Carlo samplers using probabilistic programming][74], Workshop on Approximate Inference;
  • Monitoring the errors of discriminative models with probabilistic programming, Workshop on Reliable Machine Learning in the Wild.
  • October 2016: Invited talk on probabilistic programming at the [Uncertainty in Computation Workshop][75], hosted by the Simons Institute for the Theory of Computing in Berkeley, California.
  • September 2016: Invited tutorial on probabilistic programming at the [Machine Learning for Signal Processing Workshop][76] in Salerno, Italy.
  • September 2016: We are offering an MIT course, [9\.S915, Introduction to Probabilistic Programming][77].
  • May 2016: [BayesDB 0.1.8][78] is now available.
  • April 2016: We will be participating in the [Healthy Birth, Growth, and Development Knowledge Integration][79] convening in Seattle, organized by the [Bill and Melinda Gates Foundation][80].
  • March 2016: Video of our MIT Media Lab talk on [Probabilistic Programming for Augmented Intelligence][81] is now available.
  • March 2016: [BayesDB 0.1.7][78] is now available.
  • February 2016: [BayesDB 0.1.6][78] is now available.
  • January 2016: [Venture Alpha 0.5][82] is now available.
  • January 2016: Empirical Systems, Inc., our first spin-out company, has secured seed financing from backers including Matrix Partners and Flybridge Capital.
  • January 2016: We are offering an MIT IAP course on probabilistic programming for data analysis. It will cover BayesDB, Picture, some Venture, and material on AI from our talk at the MIT [ILP Seminar on Minds, Machines, and Management][83] in Vienna (March 2015).
  • January 2016: [BayesDB 0.1.5][78] is now available.
  • January 2016: We are delighted to be presenting on BayesDB at the [Northeast Database Day][84] hosted by the CSAIL Big Data Initiative.
  • December 2015: Videos of our Strange Loop Future Programming Workshop talks on probabilistic programming and on BayesDB are now available.
  • Accessibility
  • [7]: https://proceedings.mlr.press/v151/saad22a.html
  • [8]: https://news.mit.edu/2022/estimating-informativeness-data-0425
  • [9]: https://papers.nips.cc/paper/2021/file/4fc66104f8ada6257fa55f29a2a567c7-Paper.pdf
  • [10]: https://news.mit.edu/2021/probablistic-programming-machine-vision-1208
  • [11]: https://dl.acm.org/doi/10.1145/3453483.3454078
  • [12]: https://news.mit.edu/2021/exact-symbolic-artificial-intelligence-faster-better-assessment-ai-fairness-0809
  • [13]: https://arxiv.org/abs/2007.11838
  • [14]: https://news.mit.edu/2021/system-cleans-messy-data-tables-automatically-0511
  • [15]: https://www.auai.org/uai2021/pdf/uai2021.405.pdf
  • [16]: https://arxiv.org/abs/2010.03485
  • [17]: https://popl21.sigplan.org/room/POPL-2021-venue-lafi
  • [18]: https://openreview.net/forum?id=8Itm8dQnJRc
  • [19]: https://papers.nips.cc/paper/2020/file/df3aebc649f9e3b674eeb790a4da224e-Paper.pdf
  • [20]: https://news.mit.edu/2020/building-machines-better-understand-human-goals-1214
  • [21]: https://popl21.sigplan.org/details/lafi-2021-papers/1/On-the-Automatic-Derivation-of-Importance-Samplers-from-Pairs-of-Probabilistic-Progra
  • [22]: https://popl21.sigplan.org/details/lafi-2021-papers/5/Genify-jl-Transforming-Julia-into-Gen-to-enable-programmable-inference
  • [23]: http://proceedings.mlr.press/v119/witty20a/witty20a.pdf
  • [24]: https://arxiv.org/abs/2003.03830
  • [25]: https://news.mit.edu/2020/algorithm-simulates-roll-loaded-dice-0528
  • [26]: https://www.intel.com/content/www/us/en/research/blogs/mit-fast-loaded-dice-roller.html
  • [27]: https://www.quantamagazine.org/how-and-why-computers-roll-loaded-dice-20200708/
  • [28]: https://popl20.sigplan.org/
  • [29]: https://dl.acm.org/doi/10.1145/3371087
  • [30]: https://dl.acm.org/doi/10.1145/3371104
  • [31]: http://news.mit.edu/2019/ai-programming-gen-0626
  • [32]: https://venturebeat.com/2019/06/27/mits-gen-programming-system-allows-users-to-easily-create-computer-vision-statistical-ai-and-robotics-programs/
  • [33]: https://www.zdnet.com/article/mit-were-building-on-julia-programming-language-to-open-up-ai-coding-to-novices/
  • [34]: https://news.ycombinator.com/item?id=20301352
  • [35]: https://dl.acm.org/citation.cfm?id=3314642
  • [36]: https://dl.acm.org/citation.cfm?doid=3302515.3290350
  • [37]: http://news.mit.edu/2019/nonprogrammers-data-science-0115
  • [38]: https://www.zdnet.com/article/mit-aims-to-help-data-scientists-by-having-ai-do-the-heavy-lifting/
  • [39]: http://proceedings.mlr.press/v89/saad19a.html
  • [40]: http://fsaad.mit.edu/assets/a37-saad.webm
  • [41]: https://probprog.cc/
  • [42]: https://www.youtube.com/playlist?list=PL_PW0E_Tf2qvXBEpl10Y39RULTN-ExzZQ
  • [43]: https://ww2.amstat.org/meetings/jsm/2018/onlineprogram/AbstractDetails.cfm?abstractid=330243
  • [44]: https://www.tableau.com/about/press-releases/2018/tableau-acquires-empirical-systems
  • [45]: https://d60.darpa.mil/
  • [46]: https://www.itu.int/en/ITU-T/AI/2018/Pages/default.aspx
  • [47]: https://arxiv.org/abs/1704.01087
  • [48]: https://pldi18.sigplan.org/track/mapl-2018-papers#Program
  • [49]: https://dl.acm.org/citation.cfm?id=3192409
  • [50]: https://dl.acm.org/citation.cfm?id=3192399
  • [51]: https://arxiv.org/abs/1710.06900
  • [52]: https://www.aistats.org/aistats2018/
  • [53]: https://arxiv.org/abs/1801.03612
  • [54]: https://popl18.sigplan.org/track/pps-2018
  • [56]: https://nips.cc/Conferences/2017/Schedule?showEvent=8733
  • [57]: https://arxiv.org/abs/1705.07224
  • [58]: http://www.it.uu.se/conferences/smc2017
  • [59]: http://ppaml.galois.com/wiki/wiki/SummerSchools/2017/Announcement
  • [60]: https://conferences.oreilly.com/artificial-intelligence/ai-ny/public/schedule/detail/59168
  • [61]: https://conferences.oreilly.com/artificial-intelligence/ai-ny/public/schedule/detail/59164
  • [62]: https://newsroom.intel.com/biographies/fellows/
  • [63]: https://www.eaglobal.org/talks/ai-assisted-data-analysis-for-humanitarian-causes/
  • [64]: https://arxiv.org/pdf/1611.07051.pdf
  • [65]: http://www.childrenshospital.org/research-and-innovation/research/labs/laboratories-of-cognitive-neuroscience
  • [66]: https://medium.com/berkman-klein-center/ai-advance-may-15-2017-2c36ee9d8dc8
  • [67]: https://cyber.harvard.edu/research/ai
  • [68]: http://proceedings.mlr.press/v54/saad17a/saad17a.pdf
  • [69]: https://arxiv.org/abs/1704.04977
  • [70]: http://dalimeeting.org/dali2017/
  • [71]: https://arxiv.org/abs/1612.04759
  • [72]: https://papers.nips.cc/paper/6060-a-probabilistic-programming-approach-to-probabilistic-data-analysis.pdf
  • [73]: https://arxiv.org/abs/1611.07051v2
  • [74]: https://arxiv.org/abs/1612.02161
  • [75]: https://simons.berkeley.edu/workshops/logic2016-1
  • [76]: http://mlsp2016.conwiz.dk/tutorials.htm#mansinghka
  • [77]: https://probcomp.csail.mit.edu/9.S915/
  • [78]: https://probcomp.csail.mit.edu/software/bayesdb/
  • [79]: http://www.gatesfoundation.org/What-We-Do/Global-Health/Discovery-and-Translational-Sciences#bodyregion_0_interiorarticle_0_strategysections_3_strategysubsections1e9e85ded3fb4757a1d86dc8508751d9_0_lnkHeader
  • [80]: http://www.gatesfoundation.org/
  • [81]: https://www.media.mit.edu/video/view/mansinghka-2016-03-15
  • [82]: https://probcomp.csail.mit.edu/software/venture/
  • [83]: http://ilp.mit.edu/conference.jsp?confid=97
  • [84]: http://mitdbg.github.io/nedbday/2016/