Algebraic Geometry: A Window to Machine Learning
This IPAM workshop explores how algebraic geometry can provide new theoretical insights and frameworks for understanding machine learning systems.
Funder: Institute for Pure and Applied Mathematics
Due Dates: December 9, 2026 (application deadline for financial support/fullest consideration)
Funding Amounts: Partial travel and registration support available; registration fees: $25 (graduate student), $50 (postdoc), $75 (faculty/gov/military), $100 (industry), $10 (remote)
Summary: Workshop uniting algebraic geometry and machine learning researchers to develop new theoretical frameworks for learning systems.
Key Information: Funding priority given to recent PhDs, graduate students, and early-career researchers, but all career stages may apply.
Description
This workshop, hosted by the Institute for Pure and Applied Mathematics (IPAM), aims to bridge the fields of algebraic geometry and machine learning. It will explore how algebraic and geometric structures can provide principled, rigorous frameworks for analyzing deep learning phenomena such as neural collapse, feature learning, overparameterization, and generalization. The event is designed to foster interdisciplinary collaboration and advance the theoretical understanding of modern learning systems, leveraging recent developments in computational algebra, tensor geometry, and semi-algebraic methods. The program will include talks, discussions, and a poster session, encouraging participation from both established researchers and those early in their careers.
