Just Leaf It: Accelerating Diffusion Classifiers with Hierarchical Class Pruning
Fuente:
arXiv
Guardado en:
| Autores principales: | , , , , , |
|---|---|
| Formato: | Preprint |
| Publicado: |
2024
|
| Materias: | |
| Acceso en línea: | |
| Etiquetas: |
Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
|
| _version_ | 1866909531043266560 |
|---|---|
| author | Shanbhag, Arundhati S. Moser, Brian B. Nauen, Tobias C. Frolov, Stanislav Raue, Federico Dengel, Andreas |
| author_facet | Shanbhag, Arundhati S. Moser, Brian B. Nauen, Tobias C. Frolov, Stanislav Raue, Federico Dengel, Andreas |
| contents | Diffusion models, celebrated for their generative capabilities, have recently demonstrated surprising effectiveness in image classification tasks by using Bayes' theorem. Yet, current diffusion classifiers must evaluate every label candidate for each input, creating high computational costs that impede their use in large-scale applications. To address this limitation, we propose a Hierarchical Diffusion Classifier (HDC) that exploits hierarchical label structures or well-defined parent-child relationships in the dataset. By pruning irrelevant high-level categories and refining predictions only within relevant subcategories (leaf nodes and sub-trees), HDC reduces the total number of class evaluations. As a result, HDC can speed up inference by as much as 60% while preserving and sometimes even improving classification accuracy. In summary, our work provides a tunable control mechanism between speed and precision, making diffusion-based classification more feasible for large-scale applications. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2411_12073 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | Just Leaf It: Accelerating Diffusion Classifiers with Hierarchical Class Pruning Shanbhag, Arundhati S. Moser, Brian B. Nauen, Tobias C. Frolov, Stanislav Raue, Federico Dengel, Andreas Computer Vision and Pattern Recognition Artificial Intelligence Machine Learning Diffusion models, celebrated for their generative capabilities, have recently demonstrated surprising effectiveness in image classification tasks by using Bayes' theorem. Yet, current diffusion classifiers must evaluate every label candidate for each input, creating high computational costs that impede their use in large-scale applications. To address this limitation, we propose a Hierarchical Diffusion Classifier (HDC) that exploits hierarchical label structures or well-defined parent-child relationships in the dataset. By pruning irrelevant high-level categories and refining predictions only within relevant subcategories (leaf nodes and sub-trees), HDC reduces the total number of class evaluations. As a result, HDC can speed up inference by as much as 60% while preserving and sometimes even improving classification accuracy. In summary, our work provides a tunable control mechanism between speed and precision, making diffusion-based classification more feasible for large-scale applications. |
| title | Just Leaf It: Accelerating Diffusion Classifiers with Hierarchical Class Pruning |
| topic | Computer Vision and Pattern Recognition Artificial Intelligence Machine Learning |
| url | https://arxiv.org/abs/2411.12073 |