Research map / Label-efficient and robust learning / Test-time adaptation
NOTE: Robust Continual Test-time Adaptation Against Temporal Correlation
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 |
|---|---|---|---|
| add IABN (instance-aware batch norm) to the source model (ablation) | Source model · CIFAR10-C temporally correlated | classification error (%): 42.3 -> 24.6 | Helped |
| add PBRS memory sampling on top of IABN (ablation) | IABN · CIFAR10-C temporally correlated | classification error (%): 24.6 -> 21.1 | Helped |
| add PBRS memory sampling on top of IABN (ablation) | IABN · CIFAR100-C temporally correlated | classification error (%): 54.5 -> 47.0 | Helped |
| add IABN on top of PBRS (ablation) | PBRS · CIFAR100-C temporally correlated | classification error (%): 51.7 -> 47.0 | Helped |
| replace reservoir sampling with prediction-balanced reservoir sampling (ablation) | IABN+RS · CIFAR100-C temporally correlated | classification error (%): 48.2 -> 47.0 | Helped |
| replace reservoir sampling with prediction-balanced reservoir sampling (ablation) | IABN+RS · CIFAR10-C temporally correlated | classification error (%): 20.5 -> 21.1 | No clear effect |
| replace instance normalization with IABN (ablation) | IN+PBRS · CIFAR100-C temporally correlated | classification error (%): 52.5 -> 47.0 | Helped |
| use PBRS under i.i.d. vs non-i.i.d. stream (ablation) | PBRS only · CIFAR10-C | classification error (%): 27.5 (non-iid) vs 25.8 (iid) | No clear effect |
| use test batch directly instead of PBRS when stream is i.i.d. (ablation) | NOTE · CIFAR10-C uniform stream | classification error (%): 20.1 -> 17.6 | Helped |
| switch from NOTE to BN Stats under temporal correlation | BN Stats · CIFAR10-C | classification error (%): 42.3 -> 73.4 | Hurt |
| switch from non-i.i.d. to i.i.d. stream | LAME · CIFAR10-C | classification error (%): 36.2 -> 44.1 | Hurt |
Closest papers by meaning
- Persistent Test-time Adaptation in Recurring Testing Scenarios (NeurIPS 2024 · similarity 0.91)
- Evaluation of Test-Time Adaptation Under Computational Time Constraints (ICML 2024 · similarity 0.90)
- Buffer layers for Test-Time Adaptation (NeurIPS 2025 · similarity 0.90)
- A Layer Selection Approach to Test Time Adaptation (AAAI 2025 · similarity 0.90)
- Towards Real-World Test-Time Adaptation: Tri-net Self-Training with Balanced Normalization (AAAI 2024 · similarity 0.90)
- Monitoring Risks in Test-Time Adaptation (NeurIPS 2025 · similarity 0.90)
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.