A Holistic Approach to Unifying Automatic Concept Extraction and Concept Importance Estimation

by   Thomas Fel, et al.

In recent years, concept-based approaches have emerged as some of the most promising explainability methods to help us interpret the decisions of Artificial Neural Networks (ANNs). These methods seek to discover intelligible visual 'concepts' buried within the complex patterns of ANN activations in two key steps: (1) concept extraction followed by (2) importance estimation. While these two steps are shared across methods, they all differ in their specific implementations. Here, we introduce a unifying theoretical framework that comprehensively defines and clarifies these two steps. This framework offers several advantages as it allows us: (i) to propose new evaluation metrics for comparing different concept extraction approaches; (ii) to leverage modern attribution methods and evaluation metrics to extend and systematically evaluate state-of-the-art concept-based approaches and importance estimation techniques; (iii) to derive theoretical guarantees regarding the optimality of such methods. We further leverage our framework to try to tackle a crucial question in explainability: how to efficiently identify clusters of data points that are classified based on a similar shared strategy. To illustrate these findings and to highlight the main strategies of a model, we introduce a visual representation called the strategic cluster graph. Finally, we present https://serre-lab.github.io/Lens, a dedicated website that offers a complete compilation of these visualizations for all classes of the ImageNet dataset.


page 2

page 9


CRAFT: Concept Recursive Activation FacTorization for Explainability

Attribution methods are a popular class of explainability methods that u...

Now You See Me (CME): Concept-based Model Extraction

Deep Neural Networks (DNNs) have achieved remarkable performance on a ra...

An Experimental Investigation into the Evaluation of Explainability Methods

EXplainable Artificial Intelligence (XAI) aims to help users to grasp th...

What I Cannot Predict, I Do Not Understand: A Human-Centered Evaluation Framework for Explainability Methods

A multitude of explainability methods and theoretical evaluation scores ...

Unlocking Feature Visualization for Deeper Networks with MAgnitude Constrained Optimization

Feature visualization has gained substantial popularity, particularly af...

Don't Lie to Me! Robust and Efficient Explainability with Verified Perturbation Analysis

A variety of methods have been proposed to try to explain how deep neura...

Please sign up or login with your details

Forgot password? Click here to reset