Sotabase

Career

· Senior Research Scientist, Google DeepMind2025–
· Assistant Professor, UC Davis, Department of Statistics2023–
· Postdoctoral Researcher, UC Berkeley, Simons Institute for the Theory of Computing2021–

Publications (23)

Trained Transformers Learn Linear Models In-Context
Journal of machine learning research · 2023
292
cited
Weak convergence methods for nonlinear partial differential equations
2012
117
cited
Benign Overfitting without Linearity: Neural Network Classifiers Trained by Gradient Descent for Noisy Linear Data
Annual Conference Computational Learning Theory · 2022
89
cited
Agnostic Learning of a Single Neuron with Gradient Descent
Neural Information Processing Systems · 2020
65
cited
Implicit Bias in Leaky ReLU Networks Trained on High-Dimensional Data
International Conference on Learning Representations · 2022
49
cited
Random Feature Amplification: Feature Learning and Generalization in Neural Networks
Journal of machine learning research · 2022
33
cited
Algorithm-Dependent Generalization Bounds for Overparameterized Deep Residual Networks
Neural Information Processing Systems · 2019
32
cited
Benign Overfitting in Linear Classifiers and Leaky ReLU Networks from KKT Conditions for Margin Maximization
Annual Conference Computational Learning Theory · 2023
30
cited
Proxy Convexity: A Unified Framework for the Analysis of Neural Networks Trained by Gradient Descent
Neural Information Processing Systems · 2021
29
cited
Provable Generalization of SGD-trained Neural Networks of Any Width in the Presence of Adversarial Label Noise
International Conference on Machine Learning · 2021
22
cited
Self-training Converts Weak Learners to Strong Learners in Mixture Models
International Conference on Artificial Intelligence and Statistics · 2021
22
cited
The Double-Edged Sword of Implicit Bias: Generalization vs. Robustness in ReLU Networks
Neural Information Processing Systems · 2023
20
cited
Provable Robustness of Adversarial Training for Learning Halfspaces with Noise
International Conference on Machine Learning · 2021
14
cited
Agnostic Learning of Halfspaces with Gradient Descent via Soft Margins
International Conference on Machine Learning · 2020
13
cited
A lower bound for $p_c$ in range-$R$ bond percolation in two and three dimensions
2016
8
cited
The Effect of SGD Batch Size on Autoencoder Learning: Sparsity, Sharpness, and Feature Learning
arXiv.org · 2023
4
cited
Hemodynamic latency is associated with reduced intelligence across the lifespan: an fMRI DCM study of aging, cerebrovascular integrity, and cognitive ability
Brain Structure and Function · 2020
3
cited
Benign Overfitting and the Geometry of the Ridge Regression Solution in Binary Classification
arXiv.org · 2025
1
cited
Gradient dynamics of single-neuron autoencoders on orthogonal data
2022
Hemodynamic latency is associated with reduced intelligence: an fMRI DCM study of aging, cerebrovascular integrity, and cognitive ability
2020
Sotabase
Spencer Frei | Researcher Profile | Sotabase | Sotabase