Research map / Label-efficient and robust learning / Test-time adaptation
CONGA:Confidence-and-Gradient-Aware Learning Rate Schedule for Test Time Adaptation
Lv Shaoran, Xinyao Li, Jingjing Li
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (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 |
|---|---|---|---|
| add CONGA schedule | Tent (ViT-Base) · ImageNet-C standard TTA | accuracy (%): 61.1 -> 64.3 | Improved |
| add CONGA schedule | Tent (ResNet-50) · ImageNet-C standard TTA | accuracy (%): 43.2 -> 45.9 | Improved |
| add CONGA schedule | DeYO (ViT-Base) · ImageNet-C standard TTA | accuracy (%): 66.2 -> 68.5 | Improved |
| add CONGA schedule | EATA (ResNet-50) · ImageNet-C CTTA | accuracy (%): 41.8 -> 42.2 | Improved |
| add CONGA schedule | Tent (ViT-Base) · ImageNet-C CTTA | accuracy (%): 62.4 -> 62.6 | No clear effect |
| add CONGA schedule | Reservoir (ViT-Base) · ImageNet-C CTTA | accuracy (%): 66.0 -> 66.6 | Improved |
| add CONGA schedule | DeYO (ViT-Base) · ImageNet-R | accuracy (%): 66.1 -> 68.3 | Improved |
| use cosine decay instead of constant LR (ablation) | Tent (ViT-Base) · ImageNet-Sketch | accuracy (%): 49.1 -> 50.1 | Improved |
| use exponential decay instead of constant LR (ablation) | Tent (ViT-Base) · ImageNet-Sketch | accuracy (%): 49.1 -> 48.4 | Hurt |
| use LR decay instead of LR warmup (ablation) | FOA with CONGA · ImageNet-Sketch | accuracy (%) | Improved |
| add Learning Gate on top of Cycle Regulator (ablation) | Tent + CONGA (ViT) · ImageNet-Sketch | accuracy (%): 49.2 -> 49.5 | Improved |
| add Safety Gate on top of Cycle Regulator and Learning Gate (ablation) | Tent + CONGA (ViT) · ImageNet-Sketch | accuracy (%): 49.5 -> 50.7 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- Tent — Tent: Fully Test-Time Adaptation by Entropy Minimization (ICLR 2021) · better on 5 of 5 shared results
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- A Layer Selection Approach to Test Time Adaptation (AAAI 2025 · similarity 0.91)
- L-TTA: Lightweight Test-Time Adaptation Using a Versatile Stem Layer (NeurIPS 2024 · similarity 0.89)
- Test-Time Adaptation with Binary Feedback (ICML 2025 · similarity 0.89)
- Buffer layers for Test-Time Adaptation (NeurIPS 2025 · similarity 0.89)
- Towards Real-World Test-Time Adaptation: Tri-net Self-Training with Balanced Normalization (AAAI 2024 · similarity 0.89)
- TEA: Test-time Energy Adaptation (CVPR 2024 · 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.
