Sign In

Dan Roberts

Machine Learning @ Diffeo - theoretical physics + artificial intelligence

Dan Roberts, also known as Daniel A. Roberts, is a theoretical physicist and artificial intelligence researcher with a notable career in academia and industry. He co-founded Diffeo, a startup focused on collaborative machine intelligence, where he served as Chief Technology Officer until its acquisition by Salesforce in 2019. At Diffeo, he worked on tools that recommend novel content to users and co-organized the Knowledge Base Acceleration track for the Text Retrieval Conference (TREC).123

Roberts holds a Ph.D. in high-energy theoretical physics from MIT, where he studied holography, black holes, chaos, and computation. His academic background includes research at Duke University, the University of Cambridge, and the University of Oxford. He has also held postdoctoral positions at the Institute for Advanced Study in Princeton and has been affiliated with Facebook AI Research, Salesforce, and Sequoia Capital as an AI Fellow.123

Currently, Roberts is a researcher at OpenAI and a visiting scientist at MIT's Center for Theoretical Physics. His work bridges theoretical physics and AI, focusing on areas like neural scaling laws, robust learning, and reinforcement learning. He is also the co-author of The Principles of Deep Learning Theory.234

Related Questions

What inspired Dan Roberts to transition from theoretical physics to AI?
How did Dan Roberts' work at Diffeo influence his career in AI?
What are some of Dan Roberts' notable achievements in the field of AI?
How does Dan Roberts apply principles from theoretical physics to AI?
What is the significance of Dan Roberts being an AI Fellow at Sequoia?
Dan Roberts
Dan Roberts, photo 1
Dan Roberts, photo 2
Add to my network

Location

New York, New York