MRMCD 2025

neolium

Young being interested in all things robotic, digital or political. Avid enjoyer of caffeine and passionate despiser of generative AI. Talk to me about political theory or your infodump of the day.
Feel free to talk to me but be warned, the quality of jokes rapidly deteriorates through the evening

Any pronouns are fine.


Session

09-13
19:00
50min
How to prevent the robot uprising - an introduction to MPC based safe reinforcement learning
neolium

Machine learning in safety critical systems can be... ~a very dumb idea~ a challenge. In this talk I want to introduce a recent framework for a robust safety system for reinforcement learning in safety critical systems based on Model Predictive Control.
In short: How to train a system without blowing up the building in the process

C120 - Art. 15 GDPR