Concept-Based Abductive and Contrastive Explanations for Behaviors of Vision Models

TMLR Paper9239 Authors

Abstract: *Concept-based explanations* offer a promising approach for explaining the predictions of deep neural networks in terms of high-level, human-understandable concepts. However, existing methods either do not establish a causal connection between the concepts and model predictions or are limited in expressivity and only able to infer causal explanations involving single concepts. At the same time, the parallel line of work on *formal abductive and contrastive explanations* computes the minimal set of input features causally relevant for model outcomes but only considers low-level features such as pixels. Merging these two threads, in this work, we propose the notion of *concept-based abductive and contrastive explanations* that capture the minimal sets of high-level concepts causally relevant for model outcomes. We then present a family of algorithms that enumerate all minimal explanations while using *concept erasure* procedures to establish causal relationships. By appropriately aggregating such explanations, we are not only able to understand model predictions on individual images but also on collections of images where the model exhibits a user-specified, common *behavior*. We evaluate our approach on multiple models, datasets, and behaviors, and demonstrate its effectiveness in computing helpful, user-friendly explanations.

Submission Type: Long submission (more than 12 pages of main content)

Previous TMLR Submission Url: https://openreview.net/forum?id=JiBkFgAz9T

Changes Since Last Submission: Our previous submission (Submission Number: 8766) was desk-rejected with the comment `Modified character spacing in header`. We are not sure about the reason for this but suspect it could due to our use of the `savetrees` and `microtype` latex packages. Both of these were used for aesthetic reasons---to our eyes, the layout after the use of the `subtle` mode in `savetrees` along with `microtype`was more natural than the default layout. Note that although the use of these packages reduced the length of the main body of the paper from 16.5 pages to 16 pages, we had already declared the paper as a long submission so this reduction in length was not relevant to our decision to use these packages. However, we have removed them both from our current submission.

Assigned Action Editor: ~Bryan_Allen_Plummer1

Submission Number: 9239