Research map / Label-efficient and robust learning / Test-time adaptation
When to Retrain after Drift: A Data-Only Test of Post-Drift Data Size Sufficiency
Ren Fujiwara, Yasuko Matsubara, Yasushi Sakurai
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 |
|---|---|---|---|
| use CALIPER-chosen window instead of incremental updating | MLP · MoCap, ADWIN | MSE: 412.606 -> 7.106 | Improved |
| use CALIPER-chosen window instead of incremental updating | MLP · Dysts, ADWIN | MSE: 71.75 -> 0.432 | Improved |
| use CALIPER instead of incremental updating | Transformer · Dysts, KSWIN | MSE: 0.582 -> 0.501 | Improved |
| use CALIPER instead of incremental updating | Transformer · Automobile, ADWIN | MSE: 1.866 -> 1.948 | Hurt |
| use CALIPER instead of incremental updating | Transformer · TEP, ADWIN | MSE: 2.376 -> 1.818 | Improved |
| use CALIPER instead of incremental updating | KRR · TEP, ADWIN | MSE: 1.780 -> 2.055 | Hurt |
| use CALIPER instead of fixed window 512 | MLP · MoCap, ADWIN | MSE: 23.887 -> 7.106 | Improved |
| use CALIPER instead of best fixed window 128 (KRR) | KRR · TEP, ADWIN | MSE: 1.845 -> 2.055 | Hurt |
| use CALIPER instead of fixed window 512 (Automobile KRR) | KRR · Automobile, ADWIN | MSE: 1.902 -> 1.949 | Hurt |
| vary ESS multiplier C in {2,3,4} (ablation) | CALIPER MLP · all four datasets | test MAE | No clear effect |
| increase theta_max to 18-20 (ablation) | CALIPER MLP · MoCap | test MAE | Hurt |
| vary theta_max in {12..20} (ablation) | CALIPER MLP · TEP, Automobile, Dysts | test MAE | No clear effect |
Closest papers by meaning
- RCCDA: Adaptive Model Updates in the Presence of Concept Drift under a Constrained Resource Budget (NeurIPS 2025 · similarity 0.87)
- Identifying Predictions That Influence the Future: Detecting Performative Concept Drift in Data Streams (AAAI 2025 · similarity 0.86)
- When to retrain a machine learning model (ICML 2025 · similarity 0.86)
- Drift-Resilient TabPFN: In-Context Learning Temporal Distribution Shifts on Tabular Data (NeurIPS 2024 · similarity 0.85)
- Instance-Conditional Timescales of Decay for Non-Stationary Learning (AAAI 2024 · similarity 0.85)
- RESQUE: Quantifying Estimator to Task and Distribution Shift for Sustainable Model Reusability (AAAI 2025 · similarity 0.85)
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.
