Talk Title:

Combining Expressive Logic Programs with Machine Learning and Natural Language: Some Roadmap

Presenter:

Dr. Benjamin Grosof, DARPA

Talk Abstract:
There are exciting research opportunities to combine highly expressive logic programs together with machine learning and natural language, including via neuro-symbolic AI. This has the potential to change the trajectory of AI overall, enabling higher assurance at lower cost for a wide range of defense and civilian application tasks. We give some technical roadmap in this realm. We discuss how probabilistic uncertainty and computational scalability are central requirements.

URLs for info on the CODORD DARPA program

• as of Oct. 11, 2024 approx 4pm EDT

Solicitation:

https://sam.gov/opp/7f355f6cf3fe4827bc67cbd9e8ff76ab/view

Information Session videorecording:

https://www.youtube.com/watch?v=nX9JbNm012E

Short DARPA news item with 3-minute video:

https://www.darpa.mil/news-events/2024-09-27

Bio:

Dr. Benjamin Grosof joined DARPA in the Defense Sciences Office in September 2023. His interests in artificial intelligence foundations include meta logic programs; scalability; deeply combining logical knowledge representation and reasoning with machine learning and natural language (e.g., neuro-symbolic); explainability and interpretability; trustworthiness and critical thinking; autonomy, and human-machine interaction. His interests in AI-enabled applications include: defense intelligence, operations, planning, and information systems integration; finance, legal and policy; e-commerce and supply chain; health care, and science.

Prior to joining DARPA as a program manager, Grosof was founding CEO and chief scientist at Coherent Knowledge, an AI startup maker of open-source tools for highly explainable decision support via query answering. He has pioneered technology invention and industry standards for knowledge graphs and expressively flexible semantic rules, their acquisition from natural language, and a wide variety of applications. Previously, he was an IBM Research scientist, an MIT Sloan professor, and a technical/research executive in AI at: the Allen Institute for AI’s predecessor; Accenture; and Kyndi, a venture-backed AI startup.

Grosof received his Bachelor of Arts in applied mathematics from Harvard University and his doctorate in computer science from Stanford University.