Projects with industry and business

One day before the final presenta­tions, the com­pany be­hind the small robot pib (printable intel­ligent bot) won the Mu­nich Digi­tal Inno­vation Award. The idea of the pro­ject pib is to offer a broad user group access to hu­man­oid robots in the spirit of open source. The eu­phoria on arrival was corre­spondingly high. The stu­dents were also de­light­ed to see the robot live and in ac­tion. The pro­ject for a group of stu­dents con­sisted of teach­ing pib to play a game in which the aim was to throw a small bag filled with gran­ules or corn onto a raised plat­form.

Along­side robot­ics, artifi­cial intel­li­gence is an­other trend­ing topic that is play­ing an in­creas­ingly im­portant role in eve­ryday life. The so-called PINNs (Phys­ics In­formed Neu­ral Net­works) rep­resent a meth­od from this field, in which physi­cal in­for­mation and condi­tions can be di­rectly in­clud­ed in the calcu­lation for simu­lating physi­cal phe­nomena using neural net­works. This ap­proach is any­thing but new, but recent de­velopments in hard­ware and soft­ware made it pos­sible to use PINNs effi­cient­ly. The stu­dents were asked to simu­late elec­tro­mag­netic prob­lems with the help of PINNs.

Topics such as effi­ciency and auto­mation have be­come an essen­tial part of mod­ern com­pa­nies. On the one hand, com­panies are al­ways striv­ing to auto­mate their pro­cess­es, for ex­ample in quali­ty con­trol, and to avoid man­ual inter­ven­tions in the pro­cess. On the other hand, the algo­rithms used should also run effi­cient­ly and ro­bustly. From the point of view of math­ema­ticians or pro­grammers, this is not al­ways so easy to im­ple­ment. For ex­am­ple, a group of stu­dents had to come up with a new algo­rith­mic and math­emat­ical ap­proach to auto­mati­cally com­pare pro­duc­tion parts for errors with their com­puter de­sign.

Text: Maximilian Bauer, Elite Graduate Program "Scientific Computing"