Explaining Deep Learning Models using Causal Inference
release_q3t2jhgq4fhonoinwjaq2frv2y
by
Tanmayee Narendra, Anush Sankaran, Deepak Vijaykeerthy, Senthil
Mani
2018
Abstract
Although deep learning models have been successfully applied to a variety of
tasks, due to the millions of parameters, they are becoming increasingly opaque
and complex. In order to establish trust for their widespread commercial use,
it is important to formalize a principled framework to reason over these
models. In this work, we use ideas from causal inference to describe a general
framework to reason over CNN models. Specifically, we build a Structural Causal
Model (SCM) as an abstraction over a specific aspect of the CNN. We also
formulate a method to quantitatively rank the filters of a convolution layer
according to their counterfactual importance. We illustrate our approach with
popular CNN architectures such as LeNet5, VGG19, and ResNet32.
In text/plain
format
Archived Files and Locations
application/pdf 335.9 kB
file_vpbubzzaxnasnnolyu55xoyeda
|
arxiv.org (repository) web.archive.org (webarchive) |
1811.04376v1
access all versions, variants, and formats of this works (eg, pre-prints)