Sam Schoenholz
Artificial Intelligence · 2014–2024
A researcher in the Science corpus: 33 papers; trusts 1,315 researchers and is trusted by 457 researchers here, as inferred from the citation graph.
Subfields
- Artificial Intelligence — 458 citations into it; #1,691 of 36,961 by trust here, #1,355 by citations
- Computer Vision and Pattern Recognition — 184 citations into it; #4,642 of 21,973 by trust here, #4,771 by citations
- Optimization & Training — 163 citations into it; #514 of 2,605 by trust here, #493 by citations
- Model Architectures — 134 citations into it; #289 of 2,288 by trust here, #452 by citations
- Materials Chemistry — 106 citations into it; #152 of 11,531 by trust here, #665 by citations
- Scaling & Learning Theory — 105 citations into it; #63 of 429 by trust here, #77 by citations
- Statistical and Nonlinear Physics — 71 citations into it; #275 of 10,175 by trust here, #234 by citations
- Computational Theory and Mathematics — 53 citations into it; #563 of 12,489 by trust here, #58 by citations
Whom Sam Schoenholz trusts overall
Their strongest outbound trust, strongest first, inferred from their citations.
- J. Behler
- Yoshua Bengio
- Geoffrey E. Hinton
- Diederik P. Kingma
- Kaiming He
- Nongnuch Artrith
- Alex Krizhevsky
- Yann LeCun
- Ilya Sutskever
- Jimmy Ba
Who trusts Sam Schoenholz overall
- Greg Yang
- Lechao Xiao
- Jeffrey Pennington
- Boris Hanin
- Soufiane Hayou
- Yann Dauphin
- Zeyuan Allen-Zhu
- George Philipp
- Ryo Karakida
- Piotr A. Sokól
Papers from Sam Schoenholz's perspective
Ordered by how strongly the researchers Sam Schoenholz trusts endorse each paper. Same corpus, a different order for every researcher.
- Deep Residual Learning for Image Recognition (Computer Vision and Pattern Recognition)
- ImageNet classification with deep convolutional neural networks
- Adam: A Method for Stochastic Optimization (International Conference on Learning Representations)
- Long Short-Term Memory
- Attention is All you Need (Neural Information Processing Systems)
- Gradient-based learning applied to document recognition
- ImageNet: A large-scale hierarchical image database
- Learning Multiple Layers of Features from Tiny Images
- Very Deep Convolutional Networks for Large-Scale Image Recognition (International Conference on Learning Representations)
- ImageNet Large Scale Visual Recognition Challenge
Their papers
- Beyond the Imitation Game: Quantifying and extrapolating the capabilities of language models (2022) — 2,561 citations
- Wide neural networks of any depth evolve as linear models under gradient descent (2019) — 1,337 citations
- Deep Neural Networks as Gaussian Processes (2017) — 1,255 citations
- Adversarial Spheres (2018) — 358 citations
- Deep Information Propagation (2016) — 429 citations
- Dynamical Isometry and a Mean Field Theory of CNNs: How to Train 10, 000-Layer Vanilla Convolutional Neural Networks (2018) — 392 citations
- Intriguing Properties of Adversarial Examples (2017) — 90 citations
- Mean Field Residual Networks: On the Edge of Chaos (2017) — 219 citations
- Resurrecting the sigmoid in deep learning through dynamical isometry: theory and practice (2017) — 307 citations
- The Emergence of Spectral Universality in Deep Networks (2018) — 185 citations
- Dynamical Isometry and a Mean Field Theory of RNNs: Gating Enables Signal Propagation in Recurrent Neural Networks (2018) — 126 citations
- Deep Mean Field Theory: Layerwise Variance and Width Variation as Methods to Control Gradient Explosion (2018) — 24 citations
- Neural Tangents: Fast and Easy Infinite Neural Networks in Python (2019) — 261 citations
- Dynamical Isometry and a Mean Field Theory of LSTMs and GRUs (2019) — 44 citations
- Gradients are Not All You Need (2021) — 111 citations
- Neural Message Passing for Quantum Chemistry (2017) — 9,235 citations
- Disentangling trainability and generalization in deep learning (2019) — 35 citations
- Statistical mechanics of deep learning (2024) — 234 citations
- Prediction Errors of Molecular Machine Learning Models Lower than Hybrid DFT Error. (2017) — 593 citations
- Identifying structural flow defects in disordered solids using machine-learning methods. (2014) — 336 citations