Learning theory

The Learning theory team at Google tackles fundamental learning theory problems significant to Google.

About the team

We are dedicated to advancing the theoretical foundations of machine learning (ML). Our team has extensive expertise in a variety of areas, including learning theory, statistical learning theory, optimization, decision making under uncertainty, reinforcement learning, and theory and algorithms in general. Our mission is twofold: to foster a principled understanding of ML techniques and to leverage this knowledge in designing highly effective algorithms. Ultimately, we aim to deploy these algorithms to achieve significant impact on Google, the wider academic community, and the scientific field of ML as a whole.

Team focus summaries

Featured publications

On the convergence of Adam and Beyond
Satyen Kale
International Conference on Learning Representations (2018)
Easy Learning from Label Proportions
Robert Busa-Fekete
Travis Dick
Heejin Choi
Neurips (2023)
Multiple-policy High-confidence Policy Evaluation
Mohammad Ghavamzadeh
International Conference on Artificial Intelligence and Statistics (2023), pp. 9470-9487
Layerwise Bregman Representation Learning of Neural Networks with Applications to Knowledge Distillation
Ehsan Amid
Rohan Anil
Christopher Fifty
Transactions on Machine Learning Research, 02/23 (2023)
A Model Selection Approach for Corruption Robust Reinforcement Learning
Chen-Yu Wei
33rd International Conference on Algorithmic Learning Theory (ALT 2022) (2022)

Some of our locations