Fundamental Theory and Parallel Inference for Probabilistic Programming (10.3.1 Integrated Intelligence)
Abstract
This project was conceived with three major goals: (i) develop new probabilistic programming technology that addressed limitations of first-generation languages such as Church; (ii) demonstrate the capabilities of knowledge based AI systems written using these new probabilistic programming languages, emphasizing reflective uses of probabilistic programming; and (iii) develop mathematical theory that addresses fundamental questions associated with probabilistic programs. Over the past four years, we have accomplished all three of these goals.
Document Details
- Document Type
- Technical Report
- Publication Date
- Jul 07, 2017
- Accession Number
- AD1057510
Entities
People
- Joshua B. Tenenbaum
- Vikash Mansinghka
Organizations
- Massachusetts Institute of Technology