Kernel Regression with Infinite-Width Neural Networks on Millions of Examples

Ben Adlam, Jaehoon Lee, Shreyas Padhy, Zachary Nado, Jasper Snoek

Keywords: gaussian processes, neural tangent kernel, infinite-width neural networks

TL;DR: We enable kernel regression with infinite-width neural networks at a larger scale than was previously possible to calculate scaling laws across many orders of magnitude and achieve SotA results on protein and small molecule prediction benchmarks.

Abstract: While kernel regression remains an important practical method, its connection to neural networks as their width becomes large has initiated fresh research. These neural kernels have drastically increased performance on diverse and nonstandard data modalities but require significantly more compute, which previously limited their application to smaller datasets. We address this by massively parallelizing their computation across many GPUs. We combine this with a distributed, preconditioned conjugate gradients algorithm to enable kernel regression at a large scale (i.e. up to 5 million examples). Using this approach, we study scaling laws of several neural kernels across many orders of magnitude for the CIFAR-5m dataset. Using data augmentation to expand the original CIFAR-10 training dataset by a factor of 20, we obtain a test accuracy of 91.2\% (SotA for a pure kernel method). Finally, we explore other data modalities, obtaining results on protein and small molecule prediction tasks that are competitive with SotA methods.

Anonymous Url: I certify that there is no URL (e.g., github page) that could be used to find authors’ identity.

No Acknowledgement Section: I certify that there is no acknowledgement section in this submission for double blind review.

Code Of Ethics: I acknowledge that I and all co-authors of this work have read and commit to adhering to the ICLR Code of Ethics

Submission Guidelines: Yes

Please Choose The Closest Area That Your Submission Falls Into: General Machine Learning (ie none of the above)