Hanchen

- why Batch Normalisation works:
      Internal Covariate Shift: How Does Batch Normalization Help Optimization?, NeurIPS' 18
      Loss Landscape: Visualizing the Loss Landscape of Neural Nets, NeurIPS' 18

- SGD, Adam or ...:
      On the Convergence of Adam and Beyond, ICLR' 18 best paper

- why Flat Minima is better:
      Keeping the neural networks simple by minimizing the description length of the weights, COLT' 93
      Simplifying neural nets by discovering flat minima, NIPS' 94
      Fantastic generalization measures and where to find them, ICML' 20
      SWAD: Domain Generalization by Seeking Flat Minima, NeurIPS' 21
      Shaping the learning landscape in neural networks around wide flat minima, PNAS' 20
      Asymmetric valleys: Beyond sharp and flat local minima, NeurIPS' 19
      Entropic gradient descent algorithms and wide flat minima, ICLR' 21

- why (Self-) Distillation works:
      Why distillation helps: a statistical perspective, ICML' 21
      Towards Understanding Ensemble, Knowledge Distillation and Self-Distillation in Deep Learning, arXiv' 20

- Curated Resource:
      IBM Qiskit Summer School       Google Quantum Summer Symposium

- Curated Researchers:
      John Watrous (UWaterloo),
      John Preskill (Caltech)

- Curated Resource:
      IBM Qiskit Summer School       Google Quantum Summer Symposium

- Curated Researchers:
      Bing Ren (UCSD, 3D Genome),
      Yanxiao Zhang (West Lake)

- Curated Paper:
      Geometric and Physical Quantities Improve E(3) Equivariant Message Passing (appendix), arXiv' 21

- Curated Resource:
      Symposium on Geometry Processing       Toronto Geometry Colloquium

- Curated Researchers:
      Justin Soloman (MIT, Surface Modeling),
      Xianfeng David Gu (SUNY, Conformation),
      Xiaodong Wei (EPFL),
      Mirela Ben-Chen (Technion),
      Max Welling (UoAmsterdam, emmm),
      Taco Cohen (Qualcomm, Equivariance, Group),
      Robin Walters (NEU, PDE, Symmetry),
      Ray Wang (UCSD, Dynamical Systems),

- Curated Papers:
      Applications and Techniques for Fast Machine Learning in Science, arXiv' 21

- Curated Resource:
      Stanford Bio-X       Physics Meets ML
      CMU Scientific Machine Learning Webinar

- Curated Researchers:
      James Zou (Stanford: Biomedical, Interpretability, Robustness),
      Ron Dror (Stanford: Geometry, Biology, Geonome)

- Curated Resource:
      GAMES (in Chinese)       VALSE (in Chinese) Seminar       MIT Vision & Graphics Seminar
      TUM AI Lecture Series       Dynamic Vision and Learning @ TUM       3DGV Seminar

CSS is adapted from Jon Barron's.