AI for pu­re math

A seminar series within the TRR 358 (Paderborn / Bielefeld).

The series provides a forum for exchanging experiences on the use of AI as a tool in pure mathematics research. Members of the TRR are experimenting with these tools at varying levels; the aim is to share what works and what does not, and to learn from one another. Topics range from formalization and proof assistants (Lean) to AI-assisted coding, agentic workflows, and the foundations of neural networks and large language models. The format includes both internal reports of practical experience and talks by external speakers with deeper expertise in specific areas.

The seminar is open to all members of the TRR, regardless of prior familiarity with AI. No background beyond mathematical interest is assumed. It runs in hybrid form across the two universities; external talks are typically held via Zoom.

Core time slots are Fridays 9–11 and Mondays 10–12. Dates vary between weeks so that teaching and other recurring commitments do not exclude the same participants each time.

DateTimePlaceTopicSpeaker(s)
Fri 26 June9:15–11:00Hybrid
Paderborn D2 314
Bielefeld M4 122
Best?-Practice of AI in Pure MathLukas Kühne, Claudia Alfes, Tobias Weich
Mon June 2910:15-11:45Hybrid 
Bielefeld: Raum M4 122/126 + zoom
Advancing Geometry with AI: Multi-agent Generation of PolytopesGeorges Neaime and Alec Schmutz
Fri 3 July9–11Hybrid
Paderborn D2 314
Bielefeld M4 122
Agentic loops / agentic workflowsFabian Januszewski
Mon 13 July10–12zoomAI-Assisted Recursions for Mock Theta FunctionsMartin Raum (Chalmers University of Technology)
Fri 24 July9–11Hybrid
Paderborn D2 314
Bielefeld M4 122
Foundations of neural networks / LLMs (tbc)N.N.
beginning of winter term (from October)tbaFormalization block (3 talks): introduction (type theory and Lean) + an advanced talk, via ZoomAbhijit Aryampilly Jayanthan(UPB) and Sébastien Gouëzel (Rennes)