Research map / Label-efficient and robust learning / Test-time adaptation
Prior-free Tabular Test-time Adaptation
Rundong He, Jieming Shi
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 class prior estimation (CPE) (ablation) | PFT3A TabTransformer · Health Ins. | accuracy: 58.29 -> 74.13 | Improved |
| add class prior estimation (CPE) (ablation) | PFT3A TabTransformer · Hypertension | accuracy: 59.45 -> 63.03 | Improved |
| add class prior estimation (CPE) (ablation) | PFT3A TabTransformer · HELOC | accuracy: 60.46 -> 66.17 | Improved |
| add robust feature learning (RFL) (ablation) | PFT3A TabTransformer · ANES | accuracy: 79.33 -> 80.33 | Improved |
| add robust feature learning (RFL) (ablation) | PFT3A TabTransformer · ASSIST | accuracy: 58.39 -> 59.29 | Improved |
| add representative subspace exploration (RSE) (ablation) | PFT3A TabTransformer · Health Ins. | accuracy: 73.40 -> 74.13 | Improved |
| add representative subspace exploration (RSE) (ablation) | PFT3A TabTransformer · Hypertension | accuracy: 62.58 -> 63.03 | Improved |
| use PFT3A instead of FTAT | TabTransformer · TableShift (5 datasets avg) | accuracy: 64.25 -> 68.59 | Improved |
| use PFT3A instead of Non-Adaptation | TabTransformer · TableShift (5 datasets avg) | F1: 59.92 -> 72.65 | Improved |
| apply generic TTA methods (TENT, EATA, CoTTA, ODS) instead of no adaptation | TabTransformer · TableShift (5 datasets avg) | accuracy: 60.77 -> 57.56 | Hurt |
| use dataset-specific hyperparameters instead of fixed ASSIST-tuned configuration | PFT3A · Health Ins. | accuracy: 74.13 -> 74.34 | No clear effect |
| use PFT3A with MLP backbone instead of FTAT | MLP · TableShift (5 datasets avg) | accuracy: 64.51 -> 69.25 | Improved |
Closest papers by meaning
- TabLog: Test-Time Adaptation for Tabular Data Using Logic Rules (ICML 2024 · similarity 0.91)
- Fully Test-time Adaptation for Tabular Data (AAAI 2025 · similarity 0.90)
- Backpropagation-Free Test-Time Adaptation via Probabilistic Gaussian Alignment (NeurIPS 2025 · similarity 0.87)
- Efficient Test-time Adaptation through Candidate Verification and Divergence Shifts (NeurIPS 2026 · similarity 0.86)
- TabFSBench: Tabular Benchmark for Feature Shifts in Open Environments (ICML 2025 · similarity 0.86)
- L-TTA: Lightweight Test-Time Adaptation Using a Versatile Stem Layer (NeurIPS 2024 · similarity 0.86)
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.
