Research

Have you seen this image before? Are you curious? [source]

My main area of research is Adversarial Examples for Neural Networks. Adversarial Examples are a fundamental problem observed in Deep Learning, and potentially a huge safety and security issue. Plus, they're a lot of fun to work with!

Recently, my colleagues and I won a prize in the prestigious NeurIPS 2018 Adversarial Vision Challenge. Now I have loads of new ideas, but not enough time. So if you know about Deep Learning and are looking for a Thesis topic - don't hesitate to get in touch!


Thesis Topics

Currently I don't have any open topics. However if you are motivated and interested - just send me an email and we can work something out! Please include your CV and GitHub profile.


Publications

2020

  • Brendel, Wieland; Rauber, Jonas; Kurakin, Alexey; Papernot, Nicolas; Veliqi, Behar; Mohanty, Sharada P.; Laurent, Florian; Salathé, Marcel; Bethge, Matthias; Yu, Yaodong; Zhang, Hongyang; Xu, Susu; Zhang, Hongbao; Xie, Pengtao; Xing, Eric P.; Brunner, Thomas; Diehl, Frederik; Rony, Jérôme; Hafemann, Luiz Gustavo; Cheng, Shuyu; Dong, Yinpeng; Ning, Xuefei; Li, Wenshuo; Wang, Yu: Adversarial Vision Challenge. In: Escalera, Sergio; Herbrich, Ralf (Hrsg.): The NeurIPS '18 Competition. Springer International Publishing, 2020, 129-153 mehr…

2019

  • Brunner, Thomas; Diehl, Frederik; Knoll, Alois: Copy and Paste: A Simple But Effective Initialization Method for Black-Box Adversarial Attacks. The IEEE Conference on Computer Vision and Pattern Recognition (CVPR) Workshops, 2019 mehr…
  • Brunner, Thomas; Diehl, Frederik; Truong Le, Michael; Knoll, Alois: Leveraging Semantic Embeddings for Safety-Critical Applications. The IEEE Conference on Computer Vision and Pattern Recognition (CVPR) Workshops, 2019 mehr…
  • Brunner, Thomas; Diehl, Frederik; Truong Le, Michael; Knoll, Alois: Guessing Smart: Biased Sampling for Efficient Black-Box Adversarial Attacks. The IEEE International Conference on Computer Vision (ICCV), 2019 mehr…
  • Frederik Diehl, Thomas Brunner, Michael Truong Le and Alois Knoll: Towards Graph Pooling by Edge Contraction. ICML 2019 Workshop on Learning and Reasoning with Graph-Structured Data, 2019 mehr…
  • Frederik Diehl, Thomas Brunner, Michael Truong Le, Alois Knoll: Graph Neural Networks for Modelling Traffic Participant Interaction. IEEE Intelligent Vehicles Symposium 2019, 2019 mehr…
  • Kessler, Tobias; Bernhard, Julian; Buechel, Martin; Esterle, Klemens; Hart, Patrick; Malovetz, Daniel; Truong Le, Michael; Diehl, Frederik; Brunner, Thomas; Knoll, Alois: Bridging the Gap between Open Source Software and Vehicle Hardware for Autonomous Driving. 2019 IEEE Intelligent Vehicles Symposium (IV), IEEE, 2019 mehr…

2018

  • Michael Truong Le, Frederik Diehl, Thomas Brunner, Alois Knoll: Uncertainty Estimation for Deep Neural Object Detectors in Safety-Critical Applications. International Conference on Intelligent Transportation Systems 2018, 2018 mehr…