Research map / Label-efficient and robust learning / OOD and anomaly detection
DriftSurf: Stable-State / Reactive-State Learning under Concept Drift
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 |
|---|---|---|---|
| Use DriftSurf stable-state/reactive-state process instead of standalone drift detection (StandardDD-style detector) | DriftSurf vs StandardDD · Table 2 averages across 13 datasets | misclassification rate | Mixed |
| Greedy prediction during reactive state (switch to new model after one step of outperforming) (ablation) | DriftSurf · Table 2 averages / RCV1 and 13 datasets | misclassification rate: 0.221 vs 0.229 without greedy | Helped |
| Use DriftSurf stable/reactive process instead of standalone drift detection under 20% additive noise in SEA | DriftSurf vs MDDM · SEA20 | misclassification rate: 0.243 vs 0.289 (MDDM) | Helped |
| Use DriftSurf instead of MDDM standalone drift detection | DriftSurf vs MDDM · ELECTRICITY | misclassification rate: 0.290 vs 0.344 | Helped |
| Equal-compute divided-update setting: DriftSurf vs AUE ensemble | DriftSurf vs AUE · Table 3 (update steps divided) | misclassification rate: 0.271 vs 0.317 on CoverType; DriftSurf dominates AUE across all datasets | Helped |
| Use DriftSurf instead of AUE ensemble (equal per-step compute) | DriftSurf vs AUE · Table 2 Airline and Hyper-Slow | misclassification rate: AUE outperforms DriftSurf on Airline (0.333 vs 0.338) and Hyper-Slow (0.112 vs 0.118) | Hurt |
| Increase aggressiveness of drift detection threshold delta (lower delta) with reactive state (ablation) | DriftSurf · Figure 3 averaged across datasets | misclassification rate | No clear effect |
Closest papers by meaning
- Identifying Predictions That Influence the Future: Detecting Performative Concept Drift in Data Streams (AAAI 2025 · similarity 0.89)
- Early Concept Drift Detection via Prediction Uncertainty (AAAI 2025 · similarity 0.89)
- RCCDA: Adaptive Model Updates in the Presence of Concept Drift under a Constrained Resource Budget (NeurIPS 2025 · similarity 0.88)
- Online Boosting Adaptive Learning under Concept Drift for Multistream Classification (AAAI 2024 · similarity 0.85)
- LoRA Subtraction for Drift-Resistant Space in Exemplar-Free Continual Learning (CVPR 2025 · similarity 0.84)
- Test-time Adaptation in Non-stationary Environments via Adaptive Representation Alignment (NeurIPS 2024 · similarity 0.84)
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.