I’m a machine learning researcher at Qualcomm and finishing my PhD in Machine Learning at the University of Amsterdam where I work with Max Welling. I received a BSc in CS from Utrecht University and a MSc in AI from the University of Amsterdam (both cum laude). In 2013, I cofounded Scyfer, a company specialized in deep active learning (acquired by Qualcomm in 2017). During summer 2015 I worked on unsupervised learning of equivariant representations with Geoff Hinton at Google DeepMind. During fall 2016 / spring 2017 I spent some time at OpenAI. I received the 2017 Google PhD Fellowship. Here is my CV.

My research is focussed on learning of equivariant representations for data-efficient deep learning. Besides improving data-efficiency, “equivariance to symmetry transformations” provides one of the first rational design principles for deep neural networks, and allows them to be more easily interpreted in geometrical terms than ordinary black-box networks.

I’m very excited by the application of these methods to medical image analysis, where data-efficiency is critical. More broadly, I’m fascinated by all things related to human cognition and perception, pure mathematics, and theoretical physics.

Publications

T.S. Cohen, M. Geiger, M. Weiler, The Quite General Theory of Equivariant Convolutional Networks (under review), 2018

M. Weiler, W. Boomsma, M. Geiger, M. Welling, T.S. Cohen, 3D Steerable CNNs: Learning Rotationally Equivariant Features in Volumetric Data (under review), 2018

T.S. Cohen, M. Geiger, M. Weiler, Intertwiners between Induced Representations (with applications to the theory of equivariant neural networks), ArXiv preprint 1803.10743, 2018.
[ArXiv]

M. Winkels, T.S. Cohen, 3D G-CNNs for Pulmonary Nodule Detection. (Under review at MIDL 2018).
[pdf]

B.S. Veeling, J. Linmans, J. Winkens, T.S. Cohen, M. Welling, Rotation Equivariant CNNs for Digital Pathology. (Under review at MICCAI 2018).

J. Winkens, J. Linmans, B.S. Veeling, T.S. Cohen, M. Welling, Improved Semantic Segmentation for Histopathology using Rotation Equivariant Convolutional Networks (Under review at MIDL 2018).
[pdf]

T.S. Cohen, M. Geiger, J. Koehler, M. Welling, Spherical CNNsICLR 2018 (Best paper award).
[pdf]

E. Hoogeboom, J.W.T. Peters, T.S. Cohen, M. Welling, HexaConv. ICLR 2018.
[pdf]

T.S. Cohen, M. Geiger, J. Koehler, M. Welling, Convolutional Networks for Spherical Signals. In Principled Approaches to Deep Learning Workshop ICML 2017.
[ArXiv]

A. Eck, L.M. Zintgraf, E.F.J. de Groot, T.G.J. de Meij, T.S. Cohen, P.H.M. Savelkoul, M. Welling, A.E. Budding, Interpretation of microbiota-based diagnostics by explaining individual classifier decisions, BMC Bioinformatics, 2017.

T. Matiisen, A. Oliver, T.S. Cohen, J. Schulman, Teacher-Student Curriculum Learning. Deep Reinforcement Learning Symposium, NIPS 2017
[ArXiv]

T.S. Cohen, M. Welling, Steerable CNNs. International Conference on Learning Representations (ICLR), 2017
[pdf]

L.M. Zintgraf, T.S. Cohen, T. Adel, M. Welling, Visualizing Deep Neural Network Decisions: Prediction Difference Analysis. International Conference on Learning Representations (ICLR), 2017
[pdf]

T. Adel, T.S. Cohen, M. Caan, M. Welling, 3D Scattering Transforms for Disease Classification in Neuroimaging. Neuroimage: clinical, 2017 (accepted)

T.S. Cohen, M. Welling, Group Equivariant Convolutional Networks. Proceedings of the International Conference on Machine Learning (ICML), 2016
[pdf] [supp. mat.] [code for experiments] [G-Conv code]

L.M. Zintgraf, T.S. Cohen, M. Welling, A New Method to Visualize Deep Neural Networks. ArXiv preprint 1603.02518, 2016
[ArXiv]

T.S. Cohen, M. Welling, Harmonic Exponential Families on Manifolds. Proceedings of the International Conference on Machine Learning (ICML), 2015
[pdf] [supp. mat.]

T.S. Cohen, M. Welling, Transformation Properties of Learned Visual Representations. International Conference on Learning Representations (ICLR), 2015.
[ArXiv]

T.S. Cohen, M. Welling, Learning the Irreducible Representations of Commutative Lie Groups. Proceedings of the International Conference on Machine Learning (ICML), 2014.
[pdf] [supp. mat.]

T.S. Cohen, Learning Transformation Groups and their Invariants. Master’s thesis, University of Amsterdam, 2013. (1st place University of Amsterdam thesis prize 2014)
[pdf]