Happy, proud, and honored to have received the highest scientific award of my university, the 2022 Heinz Maier-Leibnitz Medal. Thank you @TU_Muenchen for the great environment you provide.
Stephan Günnemann
305 posts
Stephan Günnemann
@guennemann
Professor for Data Analytics and Machine Learning @ TU Munich
Joined December 2019
- An an early Christmas present, we have made the videos of our MLGS lecture (Machine Learning for Graphs and Sequential Data) available online: daml.in.tum.de/teaching/mlgs Besides ML for graphs (e.g. GNNs, node embeddings, spectral embed.) and sequences (e.g. transformers, TPPs)...
- As an overview of the research in adversarial robustness of GNNs, I have written a survey paper. It is published as a book chapter in link.springer.com/book/10.1007/9…; free preprint in.tum.de/daml/alle-news…. I hope it will be useful for researchers & practitioners interested in this field.
- If you are interested in a PhD or PostDoc position (e.g. in the fields of GNNs, ML for science, or reliable ML), just reach out to me during #NeurIPS2022. More details also online: cs.cit.tum.de/daml/offene-st…
- 11 exciting news: Our group has 10 papers at #NeurIPS2024 (incl. 1 oral + 2 spotlights) 📃🎓. And as of October 1st, I am on entrepreneurial leave 🚀. Re papers: Congrats to all co-authors. Amazing work! go.tum.de/689644 Re startup: We are hiring! pruna.ai
- I am extremely happy for and proud of my amazing team of PhD students I was lucky to advise in the last couple of years: Our group has six papers accepted at #NeurIPS2021 (in.tum.de/en/daml/all-ne…), plus further great works at ICML, ICLR, KDD, ... The applause is yours! Thank you!
- I am very happy and proud of my PhD students to have (again) 4 papers at #NeurIPS2020 - incl. one oral! The works cover the full range of our core research topics: ML for graphs & temporal data as well as reliability of ML methods, i.e. robustness & uncertainty. Here the details:
- Incorporating invariances/symmetries in neural nets, e.g. rotational invariance, is key when applying ML to real world problems like molecular property prediction, medical imaging, protein folding or LiDAR classification. Can invariances be leveraged for robustness certification?
- Super excited that three of our #ICML2023 papers advance the research field of ML for molecules. In arxiv.org/abs/2303.04791 we consider GNN surrogate models. We tackle long-range interactions via the principle of Ewald summation. With A. Kosmala, @gasteigerjo, @n_gao96 1/3
- Are you interested in a PhD position in reliable Machine Learning? Then you might want to apply at our new Konrad Zuse School of Excellence: zuseschoolrelai.de/application/ The program comes with many further benefits and opportunities. Deadline January 9th, 2023!
- Thank you! Let me use the opportunity to announce that our ML lecture has also received the teaching award of the department's student council for the best mandatory lecture in the study year 2019/2020 -- it comes with a golden punched card. Very cool!
Other classes can make the topic needlessly difficult, but what a great, on-point lecture on probabilistic interference for Machine Learning at the TU, from prof. @guennemann, really, thank you 🙏 - Happy to announce that we received the best paper award at the #ICML2020 Workshop on AI for Autonomous Driving (sites.google.com/view/aiad2020/…) for our work "Deep Representation Learning and Clustering of Traffic Scenarios" (arxiv.org/abs/2007.07740). Congrats to N. Harmening & M. Bilos.
- The best poster award of our @LogConference Munich meetup goes to: "Where Did the Gap Go? Reassessing the Long-Range Graph Benchmark" by J. Tönshoff, M. Ritzert, E. Rosenbluth, M. Grohe. lmy.de/zJLF Congrats! And thanks to all participants -- it was a great event!
- If you are interested in IT systems administration (particularly GPU computing), we are happy to receive your application: stepstone.de/stellenangebot…. Besides this position, we are continuously looking for motivated PhD candidates and PostDocs: in.tum.de/daml/offene-st… Please share.
