Research map / Label-efficient and robust learning / Test-time adaptation
Back to the Source: Diffusion-Driven Adaptation To Test-Time Corruption
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| Adapt inputs with source-trained diffusion and self-ensemble predictions before and after adaptation | ResNet-50 · ImageNet-C | accuracy: 18.7 -> 29.7 | Helped |
| Adapt inputs with source-trained diffusion and self-ensemble predictions before and after adaptation | Swin-T · ImageNet-C | accuracy: 33.1 -> 40.0 | Helped |
| Adapt inputs with source-trained diffusion and self-ensemble predictions before and after adaptation | ConvNeXt-T · ImageNet-C | accuracy: 39.3 -> 44.2 | Helped |
| Apply DDA input adaptation | ResNet-50 · ImageNet-C (other corruptions) | accuracy: 25.8 -> 29.4 | Helped |
| Apply DDA input adaptation | Swin-T · ImageNet-C (other corruptions) | accuracy: 44.2 -> 43.8 | No clear effect |
| Apply DDA input adaptation | ConvNeXt-T · ImageNet-C (other corruptions) | accuracy: 47.2 -> 46.3 | No clear effect |
| Use accelerated DEIS sampling (ablation) | DDA with ResNet-50 · ImageNet-C | accuracy: 29.7 -> 27.0 | Hurt |
| Use accelerated DEIS sampling (ablation) | DDA with ResNet-50 · ImageNet-C | runtime (s): 13.5 -> 2.4 | Helped |
| Add iterative guidance to forward+reverse diffusion (ablation) | DDA with ResNet-50 · ImageNet-C elastic corruption | accuracy: 24.5 -> 32.3 | Helped |
| Add forward noising of the input to reverse+guidance (ablation) | DDA with ResNet-50 · ImageNet-C elastic corruption | accuracy: 17.7 -> 32.3 | Helped |
| Add iterative guidance to forward+reverse diffusion (ablation) | DDA with ResNet-50 · ImageNet-C blur corruption | accuracy: 13.9 -> 12.0 | Hurt |
| Add forward noising of the input to reverse+guidance (ablation) | DDA with ResNet-50 · ImageNet-C blur corruption | accuracy: 7.6 -> 12.0 | Helped |
Closest papers by meaning
- Everything to the Synthetic: Diffusion-driven Test-time Adaptation via Synthetic-Domain Alignment (CVPR 2025 · similarity 0.91)
- ReservoirTTA: Prolonged Test-time Adaptation for Evolving and Recurring Domains (NeurIPS 2025 · similarity 0.90)
- GDA: Generalized Diffusion for Robust Test-time Adaptation (CVPR 2024 · similarity 0.89)
- Test-Time Model Adaptation with Only Forward Passes (ICML 2024 · similarity 0.88)
- Towards Real-World Test-Time Adaptation: Tri-net Self-Training with Balanced Normalization (AAAI 2024 · similarity 0.88)
- A Layer Selection Approach to Test Time Adaptation (AAAI 2025 · similarity 0.88)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.