Mathematical Sciences Foundry

  The last five years have seen AI-driven breakthroughs in research-level mathematics, prompting excitement and discussion about the future of mathematics research. On Monday, May 4 th , Helen Jenne (Whitman ’13, Data Scientist at Pacific Northwest National Laboratory) will give an overview of these developments, from AI’s role in rigorously verifying proofs to its use in discovering interesting mathematical constructions and formulating conjectures. AI for mathematics extends well beyond prompting a GPT to prove a theorem; in fact, the field has seen many success stories from smaller, targeted models. The first half of the talk will provide a survey of the current landscape, and the second half will discuss recent work applying an LLM-based approach called evolutionary program synthesis to the challenge of finding combinatorial bijections. Most examples will be drawn from graph theory and combinatorics, but no prior background in these subjects will be assumed. This...

You've used ChatGPT. You've probably been amazed, and maybe a little unsettled, by what it can do. But what is actually happening when you hit send? On Monday, April 13, 4pm, Olin 301 , Todd Hendry (Whitman '00, Partner Software Engineer at Microsoft) pulls back the curtain on large language models, explaining how they are trained on vast corpora of text, what "learning" actually means mathematically, and why building and evaluating these systems at scale is harder than it looks. We will touch on the core ideas, including transformers, loss functions, and reinforcement learning from human feedback, without requiring anything beyond calculus and curiosity. Todd will share what it is like to work at the intersection of Microsoft and OpenAI, along with the kinds of problems that researchers and engineers are actively working on.         Bio:  Todd Hendry is a Partner Software Engineer at Microsoft, where he works on large language model training and eva...