BayesDB For Data-Centric Scientific Discovery
Abstract
This work was performed under an AFRL contract for developing data-driven methods to accelerate scientific discovery and robust design. Our MIT research team transitioned 4 results we believe constitute SD2 program-level accomplishments: 1. Automated data integrity via online learning of probabilistic programs, via Fail-Fast. 2. Fast exact symbolic inference for probabilistic programs, using SPPL. 3. Gen: A scalable, general-purpose probabilistic programming platform. 4. Learning whole-genome generative models from high-dimensional wet lab data to detect unwanted host-gene interactions and predict knockout effects on sample growth rate.
Document Details
- Document Type
- Technical Report
- Publication Date
- Feb 10, 2022
- Accession Number
- AD1162978
Entities
People
- Alex Lew
- Alexey Radul
- Amanda Brower
- Cameron Freer
- Desiree Dudley
- Feras Saad
- Harish Tella
- Jonathan Rees
- Joshua Tenebaum
- Joshua Thayer
- Marco Cusumano-towner
- Rachael Paiste
- Shivam Handa
- Ulrich Schaecthle
- Veronica Weiner
- Vikash Mansinghka
- Zane Shelby
Organizations
- Massachusetts Institute of Technology