Towards robust interpretability with self-explaining neural networks

Jaakkola, Tommi; Alvarez Melis, David

This is not the latest version of this item. The latest version can be found here.

Towards robust interpretability with self-explaining neural networks

Download

Name

8003-towards-robust-interpretability-with-self-explaining-neural-networks.pdf

Description

Published version

Size

2.16 MB

Format

Unknown

Checksum (MD5)

08843120a9ad5b69f6945cbb490f157a

sword-2019-05-31T16:23:59.original.xml (130 B)

Original SWORD entry document

Author(s)

Jaakkola, Tommi

•

Alvarez Melis, David

Date Issued

2018

Citation

Jaakkola, Tommi and Alvarez Melis, David. 2018. "Towards robust interpretability with self-explaining neural networks."

Version

Final published version

Abstract

© 2018 Curran Associates Inc.All rights reserved. Most recent work on interpretability of complex machine learning models has focused on estimating a posteriori explanations for previously trained models around specific predictions. Self-explaining models where interpretability plays a key role already during learning have received much less attention. We propose three desiderata for explanations in general - explicitness, faithfulness, and stability - and show that existing methods do not satisfy them. In response, we design self-explaining models in stages, progressively generalizing linear classifiers to complex yet architecturally explicit models. Faithfulness and stability are enforced via regularization specifically tailored to such models. Experimental results across various benchmark datasets show that our framework offers a promising direction for reconciling model complexity and interpretability.

Terms of Use

Article is made available in accordance with the publisher's policy and may be subject to US copyright law. Please refer to the publisher's site for terms of use.

Persistent DSpace Link

https://hdl.handle.net/1721.1/137669.2

DOI of Published Version

https://papers.nips.cc/paper/8003-towards-robust-interpretability-with-self-explaining-neural-networks