The Power Of Probabilistic Programming with Ben Vigoda - TWiML Talk #33
Published:Jul 5, 2017 00:00
•1 min read
•Practical AI
Analysis
This article summarizes a podcast episode featuring Ben Vigoda, the founder and CEO of Gamalon. The discussion centers on probabilistic programming and its applications, particularly in structuring unstructured data. Gamalon's technology, funded by DARPA, uses Bayesian Program Synthesis to convert text into structured data. The episode delves into technical aspects like posterior distribution, sampling methods, and variational methods. The article highlights Vigoda's background, including his previous work at Lyric Semiconductor and his PhD from MIT. The focus is on the potential of probabilistic programming for various data challenges, including enterprise applications and AI assistants. The article indicates a technical discussion, suitable for those with a background in AI.
Key Takeaways
Reference
“Gamalon's first application structures unstructured data — input a paragraph or phrase of unstructured text and output a structured spreadsheet/database row or API call.”