Fanny Yang


I am an Assistant Professor in the Computer Science Department (D-INFK) at ETH Zurich. Previously I was a postdoctoral Scholar at Stanford University working with John Duchi and Percy Liang and a Junior Fellow at the Institute for Theoretical Studies at ETH Zurich working with Nicolai Meinshausen. Before that, I was a PhD student at the EECS department of UC Berkeley advised by Martin Wainwright.

Research interests

I’m generally interested in theoretically understanding and developing tools in machine learning and statistics that work well. Currently I am particularly curious about gaining theoretical understanding for the generalization properties of overparameterized models for high-dimensional data (motivated by neural networks), as well as a plethora of questions related to obtaining more trustworthy ML models, specifically distributional robustness, domain generalization and interpretability.

For the latter branch of questions I’m particularly excited about problems in the medical domain - hence if you’re facing concrete reliability issues when using ML for medical diagnostics or treatment, please don’t hesitate to ping me.

Recent talk material

  • (2023) Slides for talk at ICSDS Lisbon talk on detecting when not to trust your data, featuring recent work on quantifying hidden confounding in the presence of a small randomized control trial
  • (2023) NeurIPS Tutorial on Overparameterization and Overfitting with Vidya Muthukumar and Spencer Frei: Tutorial website, Video, Slides
  • (2023) Slides for my tutorial-style talk that I gave at the SlowDNN workshop in Abu Dhabi, Mathematics of machine learning workshop at BCAM Bilbao, and the online 1W-MINDS seminar about our work on the new bias-variance trade-off by interpolators induced by the strength of inductive bias.
  • (2022) NeurIPS Workshop “I can’t believe it’s not better”: Video, Slides
  • (2021) At ELLIS Doctoral symposium on limits of rotationally invariant kernels in high dimensions (such as NTK) and semi-supervised novelty detection slides

Recent papers

                                  1. Privacy-preserving data release leveraging optimal transport and particle gradient descent
                                    Konstantin Donhauser*, Javier Abad*, Neha Hulkund, and Fanny Yang
                                    arXiv preprint, 2024
                                  1. Hidden yet quantifiable: A lower bound for confounding strength using randomized trials
                                    Piersilvio De Bartolomeis*, Javier Abad*, Konstantin Donhauser, and Fanny Yang
                                    International Conference on Artificial Intelligence and Statistics (AISTATS), 2024
                                  2. Certified private data release for sparse Lipschitz functions
                                    Konstantin Donhauser, Johan Lokna, Amartya Sanyal, March Boedihardjo, Robert Hoenig, and Fanny Yang
                                    International Conference on Artificial Intelligence and Statistics (AISTATS), 2024
                                  3. PILLAR: How to make semi-private learning more effective
                                    Francesco Pinto, Yaxi Hu, Fanny Yang, and Amartya Sanyal
                                    IEEE Conference on Secure and Trustworthy Machine Learning (SaTML), 2024
                                  4. Tight bounds for maximum l1-margin classifiers
                                    Stefan Stojanovic, Konstantin Donhauser, and Fanny Yang
                                    Algorithmic Learning Theory (ALT), 2024
                                    1. Can semi-supervised learning use all the data effectively? A lower bound perspective
                                      Alexandru Ţifrea*, Gizem Yüce*, Amartya Sanyal, and Fanny Yang
                                      Neural Information Processing Systems (NeurIPS), Spotlight, 2023
                                    2. Margin-based sampling in high dimensions: When being active is less efficient than staying passive
                                      Alexandru Tifrea*, Jacob Clarysse*, and Fanny Yang
                                      International Conference on Machine Learning (ICML), 2023
                                    3. Strong inductive biases provably prevent harmless interpolation
                                      Michael Aerni*, Marco Milanta*, Konstantin Donhauser, and Fanny Yang
                                      International Conference on Learning Representations (ICLR), 2023
                                    4. Why adversarial training can hurt robust accuracy
                                      Jacob Clarysse, Julia Hörrmann, and Fanny Yang
                                      International Conference on Learning Representations (ICLR), 2023
                                    1. How robust accuracy suffers from certified training with convex relaxations
                                      Piersilvio De Bartolomeis, Jacob Clarysse, Amartya Sanyal, and Fanny Yang
                                      NeurIPS Workshop on empirical falsification (Long Talk) 2022
                                    2. Provable concept learning for interpretable predictions using variational inference
                                      Armeen Taeb, Nicolo Ruggeri, Carina Schnuck, and Fanny Yang
                                      ICML Workshop AI4Science 2022
                                    1. How unfair is private learning?
                                      Amartya Sanyal*, Yaxi Hu*, and Fanny Yang
                                      Conference on Uncertainty in Artificial Intelligence (UAI), Oral, 2022
                                    2. Semi-supervised novelty detection using ensembles with regularized disagreement
                                      Alexandru Țifrea, Eric Stavarache, and Fanny Yang
                                      Conference on Uncertainty in Artificial Intelligence (UAI), 2022
                                    3. Fast rates for noisy interpolation require rethinking the effects of inductive bias
                                      Konstantin Donhauser, Nicolo Ruggeri, Stefan Stojanovic, and Fanny Yang
                                      International Conference on Machine Learning (ICML), 2022
                                    4. Tight bounds for minimum l1-norm interpolation of noisy data
                                      Guillaume Wang*, Konstantin Donhauser*, and Fanny Yang
                                      International Conference on Artificial Intelligence and Statistics (AISTATS), 2022
                                      1. Self-supervised Reinforcement Learning with Independently Controllable Subgoals
                                        Andrii Zadaianchuk, Georg Martius, and Fanny Yang
                                        Conference on Robot Learning (CoRL), 2021
                                      2. How rotational invariance of common kernels prevents generalization in high dimensions
                                        Konstantin Donhauser, Mingqi Wu, and Fanny Yang
                                        International Conference on Machine Learning (ICML), 2021
                                      3. Interpolation can hurt robust generalization even when there is no noise
                                        Konstantin Donhauser*, Alexandru Tifrea*, Michael Aerni, Reinhard Heckel, and Fanny Yang
                                        Neural Information Processing Systems (NeurIPS), 2021
                                        1. Understanding and Mitigating the Tradeoff between Robustness and Accuracy
                                          Aditi Raghunathan, Sang Michael Xie, Fanny Yang, John Duchi, and Percy Liang
                                          International Conference on Machine Learning (ICML), 2020
                                          1. Invariance-inducing regularization using worst-case transformations suffices to boost accuracy and spatial robustness
                                            Fanny Yang, Zuowen Wang, and Christina Heinze-Deml
                                            Neural Information Processing Systems (NeurIPS), 2019

                                                  Selected older publications

                                                  1. Early stopping for kernel boosting algorithms: A general analysis with localized complexities
                                                    Fanny Yang*, Yuting Wei*, and Martin J Wainwright
                                                    IEEE Transactions on Information Theory (IEEE IT), 2019
                                                    Neural Information Processing Systems (NeurIPS), Spotlight, 2017
                                                  2. A framework for multi-A(rmed)/B(andit) testing with online fdr control
                                                    Fanny Yang, Aaditya Ramdas, Kevin G Jamieson, and Martin J Wainwright
                                                    Neural Information Processing Systems (NeurIPS), Spotlight, 2017
                                                  3. Statistical and computational guarantees for the Baum-Welch algorithm
                                                    Fanny Yang, Sivaraman Balakrishnan, and Martin J Wainwright
                                                    Journal for Marchine Learning Research (JMLR), 2017
                                                  4. Phase Retrieval via Structured Modulation in Paley-Wiener Spaces
                                                    F Yang, V Pohl, and H Boche
                                                    International Conference on Sampling Theory and Applications 2013

                                                  Short C.V.

                                                  01/2020 - present Assistant Professor, ETH Zurich
                                                  04/2019 - 12/2019 Postdoctoral Fellow, Stanford University
                                                  09/2018 - 09/2019 Junior Fellow (Postdoc), ETH Zurich
                                                  06/2017 - 02/2018 Applied Scientist Intern, Amazon AWS, Palo Alto
                                                  08/2013 - 08/2018 PhD, UC Berkeley
                                                  10/2010 - 08/2013 M. Sc., Technical University Munich (TUM)
                                                  10/2007 - 10/2010 B.Sc., Karlsruhe Institute of Technology (KIT)


                                                  The best way to reach me is via e-mail at fan.yang (at) however note that I cannot respond to most requests although I try to answer all research-related messages.

                                                  Meetings take place in my office: CAB G19.1
                                                  Note: Enter the north side of the CAB building and walk up to G floor
                                                  The G floor is not connected, hence it’d be a pity if you reach it in the wrong section


                                                  In my free time I play the violin, more extensively earlier. Here’s a few recordings:

                                                  2023 Hottingersaal, Zurich with Adrien Luecker, Carmen Weber: Schubert B-Flat Major Trio (Audio)
                                                  2015 Berkeley Hertz Hall with Adam Bloniarz: Brahms violin sonata 3 & Debussy violin sonata (Video)
                                                  2013 Alte Hofkapelle der Residenz Munich with Kammerorchester L’Estro: Vivalidi Winter (Video)
                                                  2010 Karlsruhe Konzerthaus, solo violin with Frank-Michael Guthmann and the KIT Sinfonieorchester:
                                                  Brahms Double Concerto (Audio) Mvt 1, Mvt 2, Mvt 3
                                                  -2005 in Stuttgart: Ysaye Violin Sonata 3 (Audio), Ysaye Violin Sonata 4 (Audio),
                                                  Pablo de Sarasate Carmen Fantasy (Audio)

                                                  I also like to take photographs, now primarily with my Fuji X-T3, 35mm f1.4.