Hi! PARIS is pleased to propose an exceptional seminar of Reinhard Heckel.
Thanks to Aymeric Dieuleveut, Hi! PARIS Fellowship holder 2021 and Assistant Professor in Statistics at École polytechnique, we have the pleasure to welcome Reinhard Heckel for an exceptional seminar.
Wednesday 06 April 2022, 2.00-3.30 PM
Telecom Paris, Room 0A128
Measuring and Enhancing Robustness in Deep Learning Based Compressive Sensing
Deep-learning based algorithms outperform traditional, handcrafted algorithms for reconstructing images from few and noisy measurements. However, neural networks may be sensitive to small, yet adversarially-selected perturbations, may perform poorly under distribution shifts, and may fail to recover small but important features in an image.
To understand the sensitivity to such perturbations, we measured the robustness of a variety of deep network based and traditional methods. Perhaps surprisingly, in the context of accelerated magnetic resonance imaging, we find no evidence that deep learning based algorithms are less robust than classical, un-trained methods. Even for natural distribution shifts, we find that classical algorithms with a single hyper-parameter tuned on a training set compromise as much in performance than a neural network with 50 million parameters. Our results indicate that the state-of-the-art deep-learning-based image reconstruction methods provide improved performance than traditional methods without compromising robustness.
Finally, we present a test-time-training method to improve the distributional robustness of deep-learning based imaging.
- Watch the replay (to come)
Reinhard Heckel is a Rudolf Moessbauer assistant professor in the Department of Electrical and Computer Engineering (ECE) at the Technical University of Munich, and an adjunct assistant professor at Rice University, where he was an assistant professor in the ECE department from 2017-2019.
Before that, he spent one and a half years as a postdoctoral researcher in the Department of Electrical Engineering and Computer Sciences at the University of California, Berkeley, and a year in the Cognitive Computing & Computational Sciences Department at IBM Research Zurich. He completed his PhD in electrical engineering in 2014 at ETH Zurich and was a visiting PhD student at the Statistics Department at Stanford University.
Reinhard is working in the intersection of machine learning and signal/information processing with a current focus on deep networks for solving inverse problems, learning from few and noisy samples, and DNA data storage.