Machine Learning

The research of Rudolf Mößbauer Tenure Track Professor Reinhard Heckel lies in the intersection of machine learning and signal/information processing, with a focus on the following areas:

  • 1) Deep networks for imaging. Learning-based methods, and in particular deep neural networks, are starting to be used for important imaging applications, such as the newest generation of medical imaging systems and smartphones. We develop algorithms that yield higher-resolution images at shorter scan times than conventional methods, and we study their performance and limitations.
  • 2) The foundations of machine learning and information processing. We develop learning algorithms and study their theoretical foundations. We are particularly interested in learning from few and noisy examples since, even in the age of big data, training data can often be surprisingly scarce and noisy.
  • 3) DNA data storage. Because of its longevity and enormous information density, DNA is a promising storage medium for digital data. We co-developed the first robust DNA storage system, and we are conducting ongoing research toward future DNA storage systems. We used our technology for the first commercial application of DNA data storage in 2020, and we are now exploring the role DNA can play in information technologies more broadly.

Publications by the Focus Group

2023

  • Yilmaz, Fatih Furkan; Heckel, Reinhard: Test-time recalibration of conformal predictors under distribution shift based on unlabeled examples. 2023 more…

2022

  • Darestani, Mohammad; Liu, Jiayu; Heckel, Reinhard: Test-Time Training Can Close the Natural Distribution Shift Performance Gap in Deep Learning Based Compressed Sensing. , 2022 more…
  • Doricchi, Andrea; Platnich, Casey M.; Gimpel, Andreas; Horn, Friederikee; Earle, Max; Lanzavecchia, German; Cortajarena, Aitziber L.; Liz-Marzán, Luis M.; Liu, Na; Heckel, Reinhard; Grass, Robert N.; Krahne, Roman; Keyser, Ulrich F.; Garoli, Denis: Emerging Approaches to DNA Data Storage: Challenges and Prospects. ACS Nano 16 (11), 2022, 17552-17571 more…
  • Klug, Tobit; Heckel, Reinhard: Scaling Laws For Deep Learning Based Image Reconstruction. , 2022 more…
  • LeJeune, Daniel; Liu, Jiayu; Heckel, Reinhard: Monotonic Risk Relationships under Distribution Shifts for Regularized Risk Minimization. , 2022 more…
  • Lin, Kang; Heckel, Reinhard: Vision Transformers Enable Fast and Robust Accelerated MRI. Medical Imaging with Deep Learning, 2022 more…
  • Meiser, Linda C.; Gimpel, Andreas L.; Deshpande, Tejas; Libort, Gabriela; Chen, Weida D.; Heckel, Reinhard; Nguyen, Bichlien H.; Strauss, Karin; Stark, Wendelin J.; Grass, Robert N.: Information decay and enzymatic information recovery for DNA data storage. Communications Biology 5 (1), 2022 more…
  • Yilmaz, Fatih Furkan; Heckel, Reinhard: Regularization-wise double descent: Why it occurs and how to eliminate it. , 2022 more…

2021

  • Darestani, Mohammad Zalbagi; Chaudhari, Akshay S.; Heckel, Reinhard: Measuring Robustness in Deep Learning Based Compressive Sensing. , 2021 more…
  • Donhauser, Konstantin; Ţifrea, Alexandru; Aerni, Michael; Heckel, Reinhard; Yang, Fanny: Interpolation can hurt robust generalization even when there is no noise. , 2021 more…
  • Heckel, Reinhard: Provable Continual Learning via Sketched Jacobian Approximations. , 2021 more…
  • Huang, Wen; Hand, Paul; Heckel, Reinhard; Voroninski, Vladislav: A Provably Convergent Scheme for Compressive Sensing Under Random Generative Priors. Journal of Fourier Analysis and Applications 27 (2), 2021 more…
  • Levick, Kel; Heckel, Reinhard; Shomorony, Ilan: Achieving the Capacity of a DNA Storage Channel with Linear Coding Schemes. , 2021 more…
  • Rey, Samuel; Segarra, Santiago; Heckel, Reinhard; Marques, Antonio G.: Untrained Graph Neural Networks for Denoising. arXiv, 2021 more…
  • Shomorony, Ilan; Heckel, Reinhard: DNA-Based Storage: Models and Fundamental Limits. IEEE Transactions on Information Theory 67 (6), 2021, 3675-3689 more…
  • Zalbagi Darestani, Mohammad; Heckel, Reinhard: Accelerated MRI With Un-Trained Neural Networks. IEEE Transactions on Computational Imaging 7, 2021, 724-733 more…

2020

  • Dai, Zhenwei; Desai, Aditya; Heckel, Reinhard; Shrivastava, Anshumali: Active Sampling Count Sketch (ASCS) for Online Sparse Estimation of a Trillion Scale Covariance Matrix. , 2020 more…
  • Heckel, Reinhard; Yilmaz, Fatih Furkan: Early Stopping in Deep Networks: Double Descent and How to Eliminate it. , 2020 more…