Our software can be accessed via the Open Probabilistic Programming Stack. The core components of the stack are:

  • BayesDB, a probabilistic programming platform for AI-assisted data science.

    • CrossCat, a domain-general Bayesian method for analyzing high-dimensional, heterogeneously-typed data tables.

    • CGPM, a library of composable probabilistic models for probabilistic data analysis.

  • Venture, a prototype general-purpose probabilistic computing platform.

  • Gen, a probabilistic meta-programming platform and high-performance run-time system suitable for production engineering in probabilistic robotics.

Video Talks + Podcasts

2018-12-04. Engineering and Reverse-Engineering Intelligence Using Probabilistic Programs, Program Induction, and Deep Learning – Joshua Tenenbaum and Vikash Mansinghka at NIPS 2017.

2017-06-04. AI Assisted Data Analysis for Humanitarian Causes – Vikash Mansinghka at Effective Altruism (EA) Global, Boston.

2016-12-01: Introducing model-based thinking into AI systems – Vikash Mansinghka and Ben Lorica on The O’Reilly Data Show Podcast.

2016-10-06: Probabilistic Programming for Augmented Intelligence – Vikash Mansinghka at the Simons Institute for Theory of Computing.

2016-03-15: Probabilistic Programming for Augmented Intelligence – Vikash Mansinghka at the MIT Media Lab.

2015-10-28: BayesDB: Query the Probable Implications of Data – Marco Cusumano-Towner and Feras Saad at the Future Programming Workshop.

2015-09-24: An Overview of Probabilistic Programming – Vikash Mansighka at Strange Loop.

2015-09-24: BayesDB: Query the Probable Implications of Data – Richard Tibbetts at Strange Loop.

2015-02-17: A Survey of Probabilistic Programming – Vikash Mansinghka at Columbia Data Science

Reading List

Papers, books, and essays serve as a shared basis of knowledge for the Probabilistic Computing Project.