A Functional Information Perspective on Model Interpretation

06/12/2022
by   Itai Gat, et al.
0

Contemporary predictive models are hard to interpret as their deep nets exploit numerous complex relations between input elements. This work suggests a theoretical framework for model interpretability by measuring the contribution of relevant features to the functional entropy of the network with respect to the input. We rely on the log-Sobolev inequality that bounds the functional entropy by the functional Fisher information with respect to the covariance of the data. This provides a principled way to measure the amount of information contribution of a subset of features to the decision function. Through extensive experiments, we show that our method surpasses existing interpretability sampling-based methods on various data signals such as image, text, and audio.

READ FULL TEXT

page 5

page 8

research
10/21/2020

Removing Bias in Multi-modal Classifiers: Regularization by Maximizing Functional Entropies

Many recent datasets contain a variety of different data modalities, for...
research
12/05/2020

Understanding Interpretability by generalized distillation in Supervised Classification

The ability to interpret decisions taken by Machine Learning (ML) models...
research
03/06/2022

Evaluation of Interpretability Methods and Perturbation Artifacts in Deep Neural Networks

The challenge of interpreting predictions from deep neural networks has ...
research
05/24/2022

Functional Network: A Novel Framework for Interpretability of Deep Neural Networks

The layered structure of deep neural networks hinders the use of numerou...
research
02/19/2021

On the Similarity between von Neumann Graph Entropy and Structural Information: Interpretation, Computation, and Applications

The von Neumann graph entropy is a measure of graph complexity based on ...
research
10/19/2020

A Framework to Learn with Interpretation

With increasingly widespread use of deep neural networks in critical dec...
research
08/25/2011

Learning from Complex Systems: On the Roles of Entropy and Fisher Information in Pairwise Isotropic Gaussian Markov Random Fields

Markov Random Field models are powerful tools for the study of complex s...

Please sign up or login with your details

Forgot password? Click here to reset