Our new X account is live! Follow @wizwand_team for updates
WorkDL logo mark

Beyond Trivial Counterfactual Explanations with Diverse Valuable Explanations

About

Explainability for machine learning models has gained considerable attention within the research community given the importance of deploying more reliable machine-learning systems. In computer vision applications, generative counterfactual methods indicate how to perturb a model's input to change its prediction, providing details about the model's decision-making. Current methods tend to generate trivial counterfactuals about a model's decisions, as they often suggest to exaggerate or remove the presence of the attribute being classified. For the machine learning practitioner, these types of counterfactuals offer little value, since they provide no new information about undesired model or data biases. In this work, we identify the problem of trivial counterfactual generation and we propose DiVE to alleviate it. DiVE learns a perturbation in a disentangled latent space that is constrained using a diversity-enforcing loss to uncover multiple valuable explanations about the model's prediction. Further, we introduce a mechanism to prevent the model from producing trivial explanations. Experiments on CelebA and Synbols demonstrate that our model improves the success rate of producing high-quality valuable explanations when compared to previous state-of-the-art methods. Code is available at https://github.com/ElementAI/beyond-trivial-explanations.

Pau Rodriguez, Massimo Caccia, Alexandre Lacoste, Lee Zamparo, Issam Laradji, Laurent Charlin, David Vazquez• 2021

Related benchmarks

TaskDatasetResultRank
Visual Counterfactual Explanation (Age)CelebA Standard
FID33.8
11
Visual Counterfactual Explanation (Smile)CelebA Standard
FID29.4
11
Visual Counterfactual Explanation (Age)CelebA-HQ
FID107.5
9
Visual Counterfactual Explanation (Smile)CelebA-HQ
FID107
9
Counterfactual Visual Explanation (Age attribute)CelebA (test)
FID33.8
6
Counterfactual Visual Explanation (Smile attribute)CelebA (test)
FID29.4
6
Counterfactual Explanation (Age)CelebA-HQ (test)
FID107.5
5
Counterfactual Explanation (Smile)CelebA-HQ (test)
FID109
5
Showing 8 of 8 rows

Other info

Follow for update