On Background Bias of Post-Hoc Concept Embeddings in Computer Vision DNNs
Fuente:
arXiv
Enregistré dans:
| Auteurs principaux: | , , , , , , , |
|---|---|
| Format: | Preprint |
| Publié: |
2025
|
| Sujets: | |
| Accès en ligne: | |
| Tags: |
Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
|
| _version_ | 1866910909267443712 |
|---|---|
| author | Schwalbe, Gesina Mikriukov, Georgii Heinert, Edgar Gerolymatos, Stavros Keser, Mert Knoll, Alois Rottmann, Matthias Mütze, Annika |
| author_facet | Schwalbe, Gesina Mikriukov, Georgii Heinert, Edgar Gerolymatos, Stavros Keser, Mert Knoll, Alois Rottmann, Matthias Mütze, Annika |
| contents | The thriving research field of concept-based explainable artificial intelligence (C-XAI) investigates how human-interpretable semantic concepts embed in the latent spaces of deep neural networks (DNNs). Post-hoc approaches therein use a set of examples to specify a concept, and determine its embeddings in DNN latent space using data driven techniques. This proved useful to uncover biases between different target (foreground or concept) classes. However, given that the background is mostly uncontrolled during training, an important question has been left unattended so far: Are/to what extent are state-of-the-art, data-driven post-hoc C-XAI approaches themselves prone to biases with respect to their backgrounds? E.g., wild animals mostly occur against vegetation backgrounds, and they seldom appear on roads. Even simple and robust C-XAI methods might abuse this shortcut for enhanced performance. A dangerous performance degradation of the concept-corner cases of animals on the road could thus remain undiscovered. This work validates and thoroughly confirms that established Net2Vec-based concept segmentation techniques frequently capture background biases, including alarming ones, such as underperformance on road scenes. For the analysis, we compare 3 established techniques from the domain of background randomization on >50 concepts from 2 datasets, and 7 diverse DNN architectures. Our results indicate that even low-cost setups can provide both valuable insight and improved background robustness. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2504_08602 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | On Background Bias of Post-Hoc Concept Embeddings in Computer Vision DNNs Schwalbe, Gesina Mikriukov, Georgii Heinert, Edgar Gerolymatos, Stavros Keser, Mert Knoll, Alois Rottmann, Matthias Mütze, Annika Computer Vision and Pattern Recognition Artificial Intelligence Machine Learning The thriving research field of concept-based explainable artificial intelligence (C-XAI) investigates how human-interpretable semantic concepts embed in the latent spaces of deep neural networks (DNNs). Post-hoc approaches therein use a set of examples to specify a concept, and determine its embeddings in DNN latent space using data driven techniques. This proved useful to uncover biases between different target (foreground or concept) classes. However, given that the background is mostly uncontrolled during training, an important question has been left unattended so far: Are/to what extent are state-of-the-art, data-driven post-hoc C-XAI approaches themselves prone to biases with respect to their backgrounds? E.g., wild animals mostly occur against vegetation backgrounds, and they seldom appear on roads. Even simple and robust C-XAI methods might abuse this shortcut for enhanced performance. A dangerous performance degradation of the concept-corner cases of animals on the road could thus remain undiscovered. This work validates and thoroughly confirms that established Net2Vec-based concept segmentation techniques frequently capture background biases, including alarming ones, such as underperformance on road scenes. For the analysis, we compare 3 established techniques from the domain of background randomization on >50 concepts from 2 datasets, and 7 diverse DNN architectures. Our results indicate that even low-cost setups can provide both valuable insight and improved background robustness. |
| title | On Background Bias of Post-Hoc Concept Embeddings in Computer Vision DNNs |
| topic | Computer Vision and Pattern Recognition Artificial Intelligence Machine Learning |
| url | https://arxiv.org/abs/2504.08602 |