Zebang Shen
ETH Zurich. Post-doctoral researcher
I am currently a post-doctoral researcher at ETH Zürich, supervised by Prof. Niao He, since the end of 2022. Prior to this role, I was a post-doctoral researcher at the University of Pennsylvania from 2019 to 2022, working under the guidance of Professors Alejandro Ribeiro and Hamed Hassani. I obtained his Bachelor’s degree and Ph.D. in 2014 and 2019, respectively, from Zhejiang University, under the supervision of Prof. Hui Qian.
I am particularly intrigued by the connection between physics and machine learning, and my current research focuses on developing neural network-based methods for solving partial differential equations using entropy dissipation principles. I am also actively involved in optimization in the probability space and stochastic optimization techniques for addressing machine learning problems.
selected publications
-
Entropy-dissipation Informed Neural Network for McKean-Vlasov Type PDEsThirty-seventh Conference on Neural Information Processing Systems, 2023
-
Self-Consistency of the Fokker Planck EquationIn Proceedings of Thirty Fifth Conference on Learning Theory , 2022
-
Sinkhorn barycenter via functional gradient descentIn Thirty-fifth Conference on Neural Information Processing Systems , 2020
-
Sinkhorn natural gradient for generative models(Spotlight) Advances in Neural Information Processing Systems, 2020
-
Stochastic conditional gradient++:(non) convex minimization and continuous submodular maximizationSIAM Journal on Optimization, 2020
-
Hessian aided policy gradientIn International conference on machine learning , 2019