Saved in:
| Main Authors: | , , , |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | https://arxiv.org/abs/2512.23210 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866917179721515008 |
|---|---|
| author | Oh, Changgyoon Jeong, Jongoh Cho, Jegyeong Yoon, Kuk-Jin |
| author_facet | Oh, Changgyoon Jeong, Jongoh Cho, Jegyeong Yoon, Kuk-Jin |
| contents | Denoising diffusion probabilistic models have brought tremendous advances in generative tasks, achieving state-of-the-art performance thus far. Current diffusion model-based applications exploit the power of learned visual representations from multistep forward-backward Markovian processes for single-task prediction tasks by attaching a task-specific decoder. However, the heuristic selection of diffusion timestep features still heavily relies on empirical intuition, often leading to sub-optimal performance biased towards certain tasks. To alleviate this constraint, we investigate the significance of versatile diffusion timestep features by adaptively selecting timesteps best suited for the few-shot dense prediction task, evaluated on an arbitrary unseen task. To this end, we propose two modules: Task-aware Timestep Selection (TTS) to select ideal diffusion timesteps based on timestep-wise losses and similarity scores, and Timestep Feature Consolidation (TFC) to consolidate the selected timestep features to improve the dense predictive performance in a few-shot setting. Accompanied by our parameter-efficient fine-tuning adapter, our framework effectively achieves superiority in dense prediction performance given only a few support queries. We empirically validate our learnable timestep consolidation method on the large-scale challenging Taskonomy dataset for dense prediction, particularly for practical universal and few-shot learning scenarios. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2512_23210 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Task-oriented Learnable Diffusion Timesteps for Universal Few-shot Learning of Dense Tasks Oh, Changgyoon Jeong, Jongoh Cho, Jegyeong Yoon, Kuk-Jin Computer Vision and Pattern Recognition Denoising diffusion probabilistic models have brought tremendous advances in generative tasks, achieving state-of-the-art performance thus far. Current diffusion model-based applications exploit the power of learned visual representations from multistep forward-backward Markovian processes for single-task prediction tasks by attaching a task-specific decoder. However, the heuristic selection of diffusion timestep features still heavily relies on empirical intuition, often leading to sub-optimal performance biased towards certain tasks. To alleviate this constraint, we investigate the significance of versatile diffusion timestep features by adaptively selecting timesteps best suited for the few-shot dense prediction task, evaluated on an arbitrary unseen task. To this end, we propose two modules: Task-aware Timestep Selection (TTS) to select ideal diffusion timesteps based on timestep-wise losses and similarity scores, and Timestep Feature Consolidation (TFC) to consolidate the selected timestep features to improve the dense predictive performance in a few-shot setting. Accompanied by our parameter-efficient fine-tuning adapter, our framework effectively achieves superiority in dense prediction performance given only a few support queries. We empirically validate our learnable timestep consolidation method on the large-scale challenging Taskonomy dataset for dense prediction, particularly for practical universal and few-shot learning scenarios. |
| title | Task-oriented Learnable Diffusion Timesteps for Universal Few-shot Learning of Dense Tasks |
| topic | Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2512.23210 |