Research scientist, Google Research, Brain team.
I focus on the foundations of statistical machine learning, including theory and systems. Percy Liang was my PhD advisor at Stanford, where I was part of the statistical machine learning group.
Research scientist, Google Research, Brain team.
I focus on the foundations of statistical machine learning, including theory and systems. Percy Liang was my PhD advisor at Stanford, where I was part of the statistical machine learning group.
I created JAX together with a few colleagues in 2017. We're still working on it.
Publications and preprints, also on scholar:
The advantages of multiple classes for reducing overfitting from test set reuse
International Conference on Machine Learning (ICML), 2019
Followed by our open problem at COLT 2019.
Measuring the effects of data parallelism on neural network training
Journal of Machine Learning Research (JMLR), 2018
Supplemented by our dataset of training measurements.
Compiling machine learning programs via high-level tracing
Machine Learning and Systems (MLSys), 2018
Reports on a nascent version of JAX.
Random features for compositional kernels
arXiv preprint arXiv:1703.07872, 2017
Estimation from indirect supervision with linear moments
International Conference on Machine Learning (ICML), 2016
Principal component projection without principal component analysis
International Conference on Machine Learning (ICML), 2016
Toward deeper understanding of neural networks: the power of initialization and a dual view on expressivity
Neural Information Processing Systems (NeurIPS), 2016
Un-regularizing: approximate proximal point and faster stochastic algorithms for empirical risk minimization
International Conference on Machine Learning (ICML), 2015
Competing with the empirical risk minimizer in a single pass
Conference on Learning Theory (COLT), 2015
Simple MAP inference via low-rank relaxations
Neural Information Processing Systems (NeurIPS), 2014
Semantic parsing on Freebase from question-answer pairs
Empirical Methods in Natural Language Processing (EMNLP), 2013
Corresponds to the initial version of SEMPRE.