Nils Strodthoff is heading the Applied Machine Learning Group at the Fraunhofer Heinrich Hertz Institute, Berlin, Germany. He studied physics at Georg-August-Universität Göttingen and Imperial College London (MSc 2009) from 2005 to 2009 and received the Dr. rer. nat. degree with distinction (summa cum laude) from Technische Universität Darmstadt in 2012. Focussing on non-perturbative quantum field theory, he held postdoctoral appointments at Heidelberg University from 2012 to 2015 and Lawrence Berkeley National Laboratory, USA from 2016-2017 funded by a DFG research fellowship. He joined the machine learning group at Fraunhofer Heinrich Hertz Institute as a senior research scientist in 2017 and was appointed group leader in 2021. He authored more than 35 peer-reviewed articles, presented at numerous international conferences, received scholarships from the Deutsche Studienstiftung and HGS-HIRe and was awarded the HGS-HIRe Excellence Award in 2011. His research interests cover self-supervised learning and quality aspect for machine learning algorithms as well as a broad range of applications ranging from biomedical signal analysis over physics to communications.

• Since January 2021, I am heading the Applied Machine Learning Group and look forward to more exciting research in applied ML.
  
  
• News article (03/06/2020) and press release (German only) (25/05/2020) covering the release of the PTB-XL ECG dataset as well as our recent efforts in automatic ECG analysis using deep learning
  
  
• Our work on self-supervised protein classification was highlighted in Sebastian Ruder's
  NLP research highlights 2019 (06/01/2020).
  
  
• For media coverage of our article on detection of myocardial infarction in ECG data see
  Algorithm matches human cardiologists in detecting heart attacks (02/07/2018)
  Neuronales Netz erkennt Herzinfarkte so gut wie erfahrene Kardiologen (12/07/2018)
  Künstliche Intelligenz erkennt Herzinfarkt so gut wie Facharzt (13/08/2018)

• applications of machine learning (ML)
  • for biomedical sensor data analysis [1,11-14,17]: ECG [1,11-13,17], EEG, EIT[14],...
  • for protein/peptide data analysis [7,9]: protein classification [7], MHC binding affinity prediction [9]
  • in physics [4,8,10,15]: generative models for statistical physics and quantum field theory [4,8,15], representation learning [10]
  • in physical-layer communications [2,3]: HARQ feedback prediction [2,3], joint source and channel coding,...


• self-supervised learning
• NLP methods for protein data [7,9]
• Contrastive learning for ECG data [17]


• aspects of quality criteria for ML algorithms [5]:
  • interpretability: PredDiff/interactions[16], applications [1,7,10,13]
  • robustness: against input perturbations via stability training [6], distribution shift, ...
  • uncertainty quantification: applications [13]
  • data quality [12]

• [17] Temesgen Mehari, and Nils Strodthoff. Self-supervised representation learning from 12-lead ECG data. arXiv preprint 2103.12676, 2021. URL arXiv Code.
• [16] Stefan Blücher, and Nils Strodthoff. PredDiff: Explanations and Interactions from Conditional Expectations. arXiv preprint 2102.13519, 2021. URL arXiv Code.
• [15] Johanna Vielhaben, and Nils Strodthoff. Generative Neural Samplers for the Quantum Heisenberg Chain. arXiv preprint 2012.10264, 2021. URL arXiv Code.
• [4] Kim Nicoli, Pan Kessel, Nils Strodthoff, Wojciech Samek, Klaus-Robert Müller, and Shinichi Nakajima. Comment on "Solving Statistical Mechanics Using VANs": Introducing saVANt - VANs Enhanced by Importance and MCMC Sampling. arXiv preprint 1903.11048, 2019. URL arXiv.

• [14] Nils Strodthoff, Claas Strodthoff, Tobias Becher, Norbert Weiler, and Inéz Frerichs. Inferring respiratory and circulatory parameters from electrical impedance tomography with deep recurrent models. IEEE Journal of Biomedical and Health Informatics (to appear), 2021. URL Publisher arXiv.
• [13] Nils Strodthoff, Patrick Wagner, Tobias Schaeffter, and Wojciech Samek. Deep Learning for ECG Analysis: Benchmarks and Insights from PTB-XL. IEEE Journal of Biomedical and Health Informatics (to appear), 2020. URL Publisher arXiv Code.
• [9] Johanna Vielhaben, Markus Wenzel, Wojciech Samek, and Nils Strodthoff. USMPep: Universal Sequence Models for Major Histocompatibility Complex Binding Affinity Prediction. BMC Bioinformatics 21, 279, 2020. URL Publisher bioRxiv Code.
• [10] Stefan Bluecher, Lukas Kades, Jan M. Pawlowski, Nils Strodthoff, Julian M. Urban. Towards Novel Insights in Lattice Field Theory with Explainable Machine Learning. Phys. Rev. D 101, 094507, 2020. URL Publisher arXiv.
• [12] Patrick Wagner, Nils Strodthoff, Ralf-Dieter Bousseljot, Dieter Kreiseler, Fatima I. Lunze. Wojciech Samek, and Tobias Schaeffter. PTB-XL, a large publicly available electrocardiography dataset. Scientific Data 7, 154, 2020. URL Publisher.
• [8] Kim Nicoli, Shinichi Nakajima, Nils Strodthoff, Wojciech Samek, Klaus-Robert Müller, and Pan Kessel. Asymptotically Unbiased Estimation of Physical Observables with Neural Samplers. Phys. Rev. E 101, 023304, 2020. URL arXiv Publisher.
• [7] Nils Strodthoff, Patrick Wagner, Markus Wenzel, and Wojciech Samek. UDSMProt: universal deep sequence models for protein classification. Bioinformatics 36, no. 8, 2401-2409, 2020. URL bioRxiv Publisher Code.
• [3] Nils Strodthoff, Barış Göktepe, Thomas Schierl, Cornelius Hellge, and Wojciech Samek. Enhanced Machine Learning Techniques for Early HARQ Feedback Prediction in 5G. IEEE Journal on Selected Areas in Communications 37, no. 11, 2573-2587, 2019. URL Publisher arXiv.
• [1] Nils Strodthoff, and Claas Strodthoff. Detecting and interpreting myocardial infarction using fully convolutional neural networks. Physiological Measurement 40, no. 1, 015001, 2019. URL Publisher arXiv.

• [6] Jan Laermann, Wojciech Samek, and Nils Strodthoff. Achieving Generalizable Robustness of Deep Neural Networks by Stability Training. Pattern Recognition. DAGM GCPR 2019. Lecture Notes in Computer Science, vol 11824, 360-373, 2019. URL Publisher arXiv.
• [2] Nils Strodthoff, Barış Göktepe, Thomas Schierl, Cornelius Hellge, and Wojciech Samek. Machine Learning Techniques for Early HARQ Feedback Prediction in 5G. IEEE Global Communications Conference Workshops (GLOBECOM), 2018. URL Publisher.

• [11] Patrick Wagner, Nils Strodthoff, Ralf-Dieter Bousseljot, Wojciech Samek, and Tobias Schaeffter. PTB-XL, a large publicly available electrocardiography dataset. PhysioNet, 2020. URL Publisher.
• [5] Nils Strodthoff, Arun Shroff, and Wojciech Samek. Aspects of Evaluation Procedures for Machine Learning Algorithms. Input Document FGAI4H-D-039, 2019. URL FGAI4H-D-039.