Attribution Mask: Filtering Out Irrelevant Features By Recursively Focusing Attention on Inputs of DNNs

02/15/2021
by   Jae-Hong Lee, et al.
0

Attribution methods calculate attributions that visually explain the predictions of deep neural networks (DNNs) by highlighting important parts of the input features. In particular, gradient-based attribution (GBA) methods are widely used because they can be easily implemented through automatic differentiation. In this study, we use the attributions that filter out irrelevant parts of the input features and then verify the effectiveness of this approach by measuring the classification accuracy of a pre-trained DNN. This is achieved by calculating and applying an attribution mask to the input features and subsequently introducing the masked features to the DNN, for which the mask is designed to recursively focus attention on the parts of the input related to the target label. The accuracy is enhanced under a certain condition, i.e., no implicit bias, which can be derived based on our theoretical insight into compressing the DNN into a single-layer neural network. We also provide Gradient * Sign-of-Input (GxSI) to obtain the attribution mask that further improves the accuracy. As an example, on CIFAR-10 that is modified using the attribution mask obtained from GxSI, we achieve the accuracy ranging from 99.8% to 99.9% without additional training.

READ FULL TEXT

page 6

page 13

page 14

page 15

research
11/16/2017

A unified view of gradient-based attribution methods for Deep Neural Networks

Understanding the flow of information in Deep Neural Networks is a chall...
research
11/15/2021

Fast Axiomatic Attribution for Neural Networks

Mitigating the dependence on spurious correlations present in the traini...
research
01/26/2022

IMACS: Image Model Attribution Comparison Summaries

Developing a suitable Deep Neural Network (DNN) often requires significa...
research
07/15/2022

Anomalous behaviour in loss-gradient based interpretability methods

Loss-gradients are used to interpret the decision making process of deep...
research
11/27/2017

DeepAPT: Nation-State APT Attribution Using End-to-End Deep Neural Networks

In recent years numerous advanced malware, aka advanced persistent threa...
research
11/30/2019

End-to-End Deep Neural Networks and Transfer Learning for Automatic Analysis of Nation-State Malware

Malware allegedly developed by nation-states, also known as advanced per...
research
04/04/2023

HarsanyiNet: Computing Accurate Shapley Values in a Single Forward Propagation

The Shapley value is widely regarded as a trustworthy attribution metric...

Please sign up or login with your details

Forgot password? Click here to reset