Share your thoughts, 1 month free Claude Pro on usSee more
WorkDL logo mark

GPyTorch: Blackbox Matrix-Matrix Gaussian Process Inference with GPU Acceleration

About

Despite advances in scalable models, the inference tools used for Gaussian processes (GPs) have yet to fully capitalize on developments in computing hardware. We present an efficient and general approach to GP inference based on Blackbox Matrix-Matrix multiplication (BBMM). BBMM inference uses a modified batched version of the conjugate gradients algorithm to derive all terms for training and inference in a single call. BBMM reduces the asymptotic complexity of exact GP inference from $O(n^3)$ to $O(n^2)$. Adapting this algorithm to scalable approximations and complex GP models simply requires a routine for efficient matrix-matrix multiplication with the kernel and its derivative. In addition, BBMM uses a specialized preconditioner to substantially speed up convergence. In experiments we show that BBMM effectively uses GPU hardware to dramatically accelerate both exact GP inference and scalable approximations. Additionally, we provide GPyTorch, a software platform for scalable GP inference via BBMM, built on PyTorch.

Jacob R. Gardner, Geoff Pleiss, David Bindel, Kilian Q. Weinberger, Andrew Gordon Wilson• 2018

Related benchmarks

TaskDatasetResultRank
End-to-end GP fit+predictSynthetic GP Data D=4
Time (ms)2.14
27
RegressionBoston no contamination
RMSE3.24
11
RegressionYacht no contamination
RMSE0.48
11
RegressionConcrete no contamination
RMSE5.14
11
RegressionPower no contamination
RMSE4.04
11
Tabular RegressionEnergy no contamination (avg. over 10 trials)
RMSE1.43
11
Tabular RegressionCarbon no contamination (avg. over 10 trials)
RMSE0.013
11
Tabular RegressionProtein no contamination (avg. over 10 trials)
RMSE356.4
11
RegressionElevators 10% outlier contamination (test)
RMSE0.004
11
RegressionConcrete 10% outlier contamination
RMSE8.46
11
Showing 10 of 28 rows

Other info

Follow for update