Concept-based Visual Counterfactual Explanations with Diffusion Models
Quick Answer
C-VCE introduces a novel diffusion framework for visual counterfactual explanations, integrating a concept bottleneck layer to enhance interpretability and robustness.
Quick Take
This method achieves improved flip rates on CelebA while producing less distorted counterfactuals compared to traditional models reliant on external classifiers, making it a practical tool for vision systems in critical applications.
Key Points
- C-VCE integrates a classifier directly into the generative model for better robustness.
- Users can toggle semantic concepts during sampling to control image modifications.
- The model achieves improved flip rates on CelebA while maintaining visual fidelity.
- A probabilistic regularizer ensures minimal and controlled edits to relevant image regions.
- C-VCE is designed for safety-critical domains, reducing reliance on external classifiers.
Paper Resources
📖 Reader Mode
~2 min readAbstract:Visual counterfactual explanations aim to answer "what minimal change to this image would flip the model's prediction?", and are increasingly important as vision models are deployed in safety-critical domains (e.g., medicine). Existing diffusion-based methods can produce realistic edits, but they rely on external classifiers that must work reliably on noisy images, which makes them fragile and hard to deploy for robust explanations. We introduce C-VCE, a new diffusion framework that builds the classifier directly into the generative model via a concept bottleneck layer, so that counterfactuals are guided by human-interpretable features (concepts) instead of a separate noise robust classifier that works with pixel-level edits. Our model lets users to toggle on/off semantic concepts during sampling, then minimally adjusts relevant image regions, while preserving the rest of the image, respecting feature correlations. To keep edits small and controlled, we add a simple probabilistic regularizer that balances "change the prediction" against "stay close to the original", plus a gradient-based mask that confines modifications to the most relevant regions. On benchmarks such as CelebA, C-VCE matches or improves flip rates while producing counterfactuals that are visually closer to the input and less distorted than baselines that depend on separate noisy-image classifiers. These properties make C-VCE a practical tool for vision systems where users need concrete "what-if" images without having to trust an additional, noise-robust classifier. More broadly, our results suggest that exposing and controlling an internal concept layer is a promising way to make powerful generative models easier to understand and safer to use.
| Comments: | Accepted at the 4th World Conference on eXplainable Artificial Intelligence (XAI 2026) |
| Subjects: | Artificial Intelligence (cs.AI); Computer Vision and Pattern Recognition (cs.CV) |
| ACM classes: | I.2.10; I.4.8; I.5.1 |
| Cite as: | arXiv:2607.22544 [cs.AI] |
| (or arXiv:2607.22544v1 [cs.AI] for this version) | |
| https://doi.org/10.48550/arXiv.2607.22544 arXiv-issued DOI via DataCite |
Submission history
From: Yassine Oueslati [view email]
[v1]
Tue, 5 May 2026 15:32:50 UTC (16,277 KB)
— Originally published at arxiv.org
Want this in your inbox every morning?
Daily brief at your local 8am — bilingual EN/中文, free.
More from arXiv cs.AI
See more →HOBA: Hierarchical On-Policy Bidding Agents for Adaptive Online Advertising
HOBA (Hierarchical On-policy Bidding Agents) is a novel hierarchical reinforcement learning framework that enhances online advertising bidding systems by improving adaptability and reducing hyperparameter tuning costs. It utilizes a for hyperparameter inference, a SARSA agent for expert model selection, and a dynamic expert pool for bid execution, achieving a +3.6% increase in target cost during large-scale deployment and outperforming state-of-the-art baselines on AuctionNet.