Recent advances in machine learning—particularly large language models, automated theorem discovery, and AI-assisted symbolic computation—promise to significantly influence the practice of pure mathematics. At the same time, algebraic geometry provides a rich source of structured data, deep conjectures, and algorithmic challenges that are well suited to modern AI/ML methods. This workshop aims to bring together researchers in algebraic geometry and experts in machine learning, automated reasoning, and computational mathematics to explore this rapidly developing interface.

A central goal of the workshop is to catalyze new collaborations between mathematicians and computer scientists by engaging participants in hands-on, problem-driven research. Rather than a traditional lecture-based format, the workshop will be organized around a “mathematical hackathon” model. On the first day, participants will propose concrete research problems at the intersection of algebraic geometry and machine learning. Participants will then self-organize into small working groups, which will meet throughout the week to actively pursue these problems.

In parallel with the working sessions, the workshop will feature tutorial-style presentations aimed at algebraic geometers who may be unfamiliar with modern AI/ML tools. These tutorials will provide practical introductions to topics such as effective prompt engineering for large language models, differences between current LLM architectures and capabilities, and specialized tools for mathematical discovery and optimization (e.g., PatternBoost, AlphaEvolve, and related systems). The emphasis will be on demystifying these tools and articulating how they can be responsibly and productively integrated into mathematical research workflows.

Another key theme of the workshop is the systematic search for patterns in mathematical data as a pathway to formulating precise conjectures and proofs in algebraic geometry. Participants will explore how ML methods can assist in detecting hidden structure, guiding experimental mathematics, and suggesting plausible generalizations—while also critically examining the limitations and risks of such approaches.

By combining collaborative problem solving, targeted tutorials, and cross-disciplinary dialogue, this workshop aims to survey the current landscape of AI-assisted mathematics, identify promising research directions, and lay the groundwork for sustained collaborations between algebraic geometers and machine learning researchers.