Have you ever wondered whether artificial intelligence can help mathematicians to commit theoretical research? There is a recent paper by Wagner (2021, [arXiv:2104.14516]) in which he proposes to find counterexamples to combinatorial conjectures by generating constructions using neural networks.
In this seminar we apply his method on several combinatorial, easy to state conjectures. Participants can also work on their own favorite combinatorial conjecture if they have one that fits into the setting. No deeper theoretical background in combinatorics or neuronal networks is required, but programming skills on at least an intermediate level are expected (preferably python).
If you consider to participate or if you have any questions, please contact me by mail in advance. There will be an initial meeting on 24.10. 14:00 next to my office (MA 509) for the organization. The schedule for further weekly or biweekly meetings will then be negotiated among the participants. If you can't make it to this initial meeting but if you would like to participate, let me know in advance.
Entry in the Vorlesungsverzeichnis