Li, MY; Grant, E; Griffiths, TL; (2023) Gaussian Process Surrogate Models for Neural Networks. In: Proceedings of the 39th Conference on Uncertainty in Artificial Intelligence (UAI 2023). (pp. pp. 1241-1252). Proceedings of Machine Learning Research: Pittsburgh, PA, USA.

Preview

Text li23c.pdf - Published Version Download (1MB) | Preview

Abstract

Not being able to understand and predict the behavior of deep learning systems makes it hard to decide what architecture and algorithm to use for a given problem. In science and engineering, modeling is a methodology used to understand complex systems whose internal processes are opaque. Modeling replaces a complex system with a simpler, more interpretable surrogate. Drawing inspiration from this, we construct a class of surrogate models for neural networks using Gaussian processes. Rather than deriving kernels for infinite neural networks, we learn kernels empirically from the naturalistic behavior of finite neural networks. We demonstrate our approach captures existing phenomena related to the spectral bias of neural networks, and then show that our surrogate models can be used to solve practical problems such as identifying which points most influence the behavior of specific neural networks and predicting which architectures and algorithms will generalize well for specific datasets.

Downloads since deposit

Loading...

13Downloads

Download activity - last month

Export as XMLJSONCSV

Loading...

Download activity - last 12 months

Export as CSVJSONXML

Loading...

Downloads by country - last 12 months

Export as JSONCSVXML

Loading...

Archive Staff Only

View Item