Prof. Dr. Reinhard Heckel


Machine Learning

Academic Career and Research Areas

Reinhard Heckel is a Rudolf Mößbauer assistant professor in the Department of Electrical and Computer Engineering at the Technical University of Munich. Before that, he was an assistant professor in the Department of Electrical and Computer Engineering at Rice University. 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, where he co-designed a recommender system currently in use. He completed his PhD in 2014 at ETH Zurich and was a visiting PhD student at the Statistics Department at Stanford University. Reinhard Heckel is working in machine learning and his current research focuses on active learning, deep generative models for learning and solving inverse problems, and DNA data storage.

Hisresearch lies in the intersection of machine learning, statistics, and signal processing. Currently he’s particularly interested in: Learning from few and noisy examples, fundamentals of deep learning, solving inverse problems with deep learning, and DNA data storage.


    • Zweiter Platz bei der AAPM (American Association of Physicists in Medicine) Grand Challenge: "Deep Learning for Inverse Problems: Sparse-View Computed Tomography Image Reconstruction" (2021)
    • In der BBC-Zukunftsserie "The Genius Behind" wird das Projekt vorgestellt: So kann man menschliches Wissen für die Ewigkeit speichern (2015)
    • ETH Zürich Medaille für herausragende Doktorarbeit (2015)
    • IBM-Preis für die erste Patentanmeldung und Erfindung (2015)
    • Early Postdoc.Mobility Stipendium des Schweizerischen Nationalfonds (2014)

    Measuring robustness in deep learning based compressive sensing M. Zalbagi Darestani, Akshay Chaudhari, and R. Heckel ICML 2021 (International Conference on Machine Learning) (long talk).


    Compressive sensing with un-trained neural networks: Gradient descent finds the smoothest approximation R. Heckel and M. Soltanolkotabi ICML 2020 (International Conference on Machine Learning).


    Low cost DNA data storage using photolithographic synthesis and advanced information reconstruction and error correction P. L. Antkowiak, J. Lietard, M. Z. Darestani, M. M. Somoza, W. J. Stark, R. Heckel & R. N. Grass Nature Communications, 2020.


    Deep decoder: Concise image representations from untrained non-convolutional networks R. Heckel and P. Hand ICLR 2019 (International Conference on Learning Representations) CODE.


    Active ranking from pairwise comparisons and when parametric assumptions don’t help R. Heckel, N. B. Shah, K. Ramchandran, and M. J. Wainwright Annals of Statistics, 2019.