6th Leibniz MMS Days
April 17 - April 19, 2023
Potsdam

Material from the World Café Tables

This page contains notes for review purposes for participants and other interested parties, rather than scripted reports.

We should not be afraid to talk in more detail about the mathematical methods than to present the respective applications very detailed.

Planning a series of Continued Education events on MMS issues
- e.g. on "How to produce good code" (understandable for colleagues, reproducible, etc.)
- online, i.e. 3-4 hours
MMS Science Slam (apart or within the scope of the MMS Days - or a mixture)
more focus on open problems (instead of results as it is the case at the MMS Days)
- Matchmaking Events (online)
- in the scope of a Science Slam
Training/information events on "How to write a successful MMS related proposal" (e.g within the Leibniz Competition)

(notes from table facilitator Torsten K"ouml;hler, WIAS)

AI is a good first guess to speed up the scientific process,
AI is considered as a "tool" in the scientific process - similar to "finite element method", ODE/PDE solver or simulations,
AI does not know the "ground truth", but giving a good first "estimate" of the problem -> result should be recalculated with "exact" methods,
be cautious in the interpretation of AI's results -> as intrinsic properties AI will "generate" also unscientific results and correlation,
be also cautious about some bias in the learning data -> check the training set as this will effect the result,
there's no standard method of AI yet as the method is quite new in the scientific community -> need for standardization and explanation of AI model parameter,
Open discussion in scientific community and society: legal liability and rights hast to be clarified -> How to use, how to attribute, who owns the data (copyright issue unclear)?

(notes from table facilitator Ron Dockhorn, IPF Dresden)