Deep Kernel Machines via the Kernel Reparametrization Trick
Abstract
While deep neural networks have achieved state-of-the-art performance on many tasks across varied domains, they still remain black boxes whose inner workings are hard to interpret and understand. In this paper, we develop a novel method for efficiently capturing the behaviour of deep neural networks using kernels. In particular, we construct a hierarchy of increasingly complex kernels that encode individual hidden layers of the network. Furthermore, we discuss how our framework motivates a novel supervised weight initialization method that discovers highly discriminative features already at initialization.
Cite
Text
Mitrovic et al. "Deep Kernel Machines via the Kernel Reparametrization Trick." International Conference on Learning Representations, 2017.Markdown
[Mitrovic et al. "Deep Kernel Machines via the Kernel Reparametrization Trick." International Conference on Learning Representations, 2017.](https://mlanthology.org/iclr/2017/mitrovic2017iclr-deep/)BibTeX
@inproceedings{mitrovic2017iclr-deep,
title = {{Deep Kernel Machines via the Kernel Reparametrization Trick}},
author = {Mitrovic, Jovana and Sejdinovic, Dino and Teh, Yee Whye},
booktitle = {International Conference on Learning Representations},
year = {2017},
url = {https://mlanthology.org/iclr/2017/mitrovic2017iclr-deep/}
}