Contrastive Corpus Attribution for Explaining Representations

by   Chris Lin, et al.

Despite the widespread use of unsupervised models, very few methods are designed to explain them. Most explanation methods explain a scalar model output. However, unsupervised models output representation vectors, the elements of which are not good candidates to explain because they lack semantic meaning. To bridge this gap, recent works defined a scalar explanation output: a dot product-based similarity in the representation space to the sample being explained (i.e., an explicand). Although this enabled explanations of unsupervised models, the interpretation of this approach can still be opaque because similarity to the explicand's representation may not be meaningful to humans. To address this, we propose contrastive corpus similarity, a novel and semantically meaningful scalar explanation output based on a reference corpus and a contrasting foil set of samples. We demonstrate that contrastive corpus similarity is compatible with many post-hoc feature attribution methods to generate COntrastive COrpus Attributions (COCOA) and quantitatively verify that features important to the corpus are identified. We showcase the utility of COCOA in two ways: (i) we draw insights by explaining augmentations of the same image in a contrastive learning setting (SimCLR); and (ii) we perform zero-shot object localization by explaining the similarity of image representations to jointly learned text representations (CLIP).


page 3

page 8

page 9

page 25

page 26

page 27


Two-Stage Holistic and Contrastive Explanation of Image Classification

The need to explain the output of a deep neural network classifier is no...

Contrastive Explanations for Explaining Model Adaptations

Many decision making systems deployed in the real world are not static -...

Explaining Latent Representations with a Corpus of Examples

Modern machine learning models are complicated. Most of them rely on con...

Towards Visually Explaining Similarity Models

We consider the problem of visually explaining similarity models, i.e., ...

RELAX: Representation Learning Explainability

Despite the significant improvements that representation learning via se...

Explaining a Series of Models by Propagating Local Feature Attributions

Pipelines involving a series of several machine learning models (e.g., s...

Please sign up or login with your details

Forgot password? Click here to reset