ML4CPS: Machine Learning For Cyber-Physical Systems ARIC Hamburg Hamburg, Germany, March 29-31, 2023 |
Conference website | https://www.hsu-hh.de/imb/en/ml4cps |
Submission link | https://easychair.org/conferences/?conf=ml4cps0 |
Submission deadline | January 20, 2023 |
Submission Short Papers | January 27, 2023 |
Cyber-physical systems can adapt to changing requirements. In combination withmachine learning, application fields such as predictive maintenance, self-optimization or fault diagnosis come into mind. One of the main prerequisites is that machine learning methods can be used by engineers.
Therefore, the 6th Machine Learning 4 CyberPhysical Systems - ML4CPS - conference offers researchers and users from various fields an exchange platform and will take place from 29th till 31st of March 2023 at the Artificial Intelligence Center Hamburg (ARIC). Hosts are the Helmut-Schmidt-University(HSU) and the Fraunhofer Institute of Optronics, System Technologies and Image Exploitation IOSB.
Submission Guidelines: Papers are chosen on a peer-review basis and accepted papers are published by Springer with a DOI number. Papers with commercial character won’t be considered. Papers must not belonger than 6-8 pages. For further information and how to submit please refer to:https://www.hsu-hh.de/imb/en/ml4cps
Papers may cover, but are not limited to the following topics:
- Integrating domain knowledge intoneural networks: This a key factor forrobust and performant neural networksin cyber-physical systems. Examples how prior knowledge can be integrated into theneural network are the network architecture, additional data from simulations or by adding constraints to the loss function.
- Certification of ML models: Performance, robustness and sensitivity of a model must be evaluated according to certain criteria. How can such criteria look like and ML models be certified by adopting them?
- Automatic protection of ML models: For cyber-physical systems it is important that ML models do not deliver unexpected results. Nevertheless, how do solutions look like that put ML models into a fall-backmode? How can it be proven that the ML model is safe to use?
- Automated Machine Learning: For the use of ML models in practice, it is essential that they can be trained, validated, and put into operation quickly. AutoML is an efficient tool for that. How can AutoML be used to fulfill multicriterial objects, like interpretability or model size in combination with performance?
All questions related to paper submissions should be emailed to ml4cps_orga@hsu-hh.de.