QPM: Discrete Optimization for Globally Interpretable Image Classification
Fuente:
arXiv
Guardado en:
| Autores principales: | , , , , |
|---|---|
| Formato: | Preprint |
| Publicado: |
2025
|
| Materias: | |
| Acceso en línea: | |
| Etiquetas: |
Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
|
| _version_ | 1866929734407946240 |
|---|---|
| author | Norrenbrock, Thomas Kaiser, Timo Biswas, Sovan Manuvinakurike, Ramesh Rosenhahn, Bodo |
| author_facet | Norrenbrock, Thomas Kaiser, Timo Biswas, Sovan Manuvinakurike, Ramesh Rosenhahn, Bodo |
| contents | Understanding the classifications of deep neural networks, e.g. used in safety-critical situations, is becoming increasingly important. While recent models can locally explain a single decision, to provide a faithful global explanation about an accurate model's general behavior is a more challenging open task. Towards that goal, we introduce the Quadratic Programming Enhanced Model (QPM), which learns globally interpretable class representations. QPM represents every class with a binary assignment of very few, typically 5, features, that are also assigned to other classes, ensuring easily comparable contrastive class representations. This compact binary assignment is found using discrete optimization based on predefined similarity measures and interpretability constraints. The resulting optimal assignment is used to fine-tune the diverse features, so that each of them becomes the shared general concept between the assigned classes. Extensive evaluations show that QPM delivers unprecedented global interpretability across small and large-scale datasets while setting the state of the art for the accuracy of interpretable models. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2502_20130 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | QPM: Discrete Optimization for Globally Interpretable Image Classification Norrenbrock, Thomas Kaiser, Timo Biswas, Sovan Manuvinakurike, Ramesh Rosenhahn, Bodo Computer Vision and Pattern Recognition Artificial Intelligence Human-Computer Interaction Machine Learning Understanding the classifications of deep neural networks, e.g. used in safety-critical situations, is becoming increasingly important. While recent models can locally explain a single decision, to provide a faithful global explanation about an accurate model's general behavior is a more challenging open task. Towards that goal, we introduce the Quadratic Programming Enhanced Model (QPM), which learns globally interpretable class representations. QPM represents every class with a binary assignment of very few, typically 5, features, that are also assigned to other classes, ensuring easily comparable contrastive class representations. This compact binary assignment is found using discrete optimization based on predefined similarity measures and interpretability constraints. The resulting optimal assignment is used to fine-tune the diverse features, so that each of them becomes the shared general concept between the assigned classes. Extensive evaluations show that QPM delivers unprecedented global interpretability across small and large-scale datasets while setting the state of the art for the accuracy of interpretable models. |
| title | QPM: Discrete Optimization for Globally Interpretable Image Classification |
| topic | Computer Vision and Pattern Recognition Artificial Intelligence Human-Computer Interaction Machine Learning |
| url | https://arxiv.org/abs/2502.20130 |