Back to Search Start Over

Enhanced Recurrent Neural Tangent Kernels for Non-Time-Series Data

Authors :
Alemohammad, Sina
Balestriero, Randall
Wang, Zichao
Baraniuk, Richard
Publication Year :
2020

Abstract

Kernels derived from deep neural networks (DNNs) in the infinite-width regime provide not only high performance in a range of machine learning tasks but also new theoretical insights into DNN training dynamics and generalization. In this paper, we extend the family of kernels associated with recurrent neural networks (RNNs), which were previously derived only for simple RNNs, to more complex architectures including bidirectional RNNs and RNNs with average pooling. We also develop a fast GPU implementation to exploit the full practical potential of the kernels. Though RNNs are typically only applied to time-series data, we demonstrate that classifiers using RNN-based kernels outperform a range of baseline methods on 90 non-time-series datasets from the UCI data repository.

Details

Database :
arXiv
Publication Type :
Report
Accession number :
edsarx.2012.04859
Document Type :
Working Paper