Research map / Label-efficient and robust learning / Recognition and long-tail
Learning with Feature Evolvable Streams
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 adaptive weighted combination of predictions from recovered old features and current features | FESL-c · Australian | accuracy: .849$\pm$.009 -> .849$\pm$.009 | No clear effect |
| Use adaptive weighted combination of predictions from recovered old features and current features | FESL-c · r.EN-FR | accuracy: .902$\pm$.004 -> .903$\pm$.003 | Helped |
| Use dynamic selection instead of weighted combination | FESL-s · Australian | accuracy: .849$\pm$.009 -> .849$\pm$.009 | No clear effect |
| Use dynamic selection instead of weighted combination | FESL-s · r.EN-FR | accuracy: .903$\pm$.003 -> .902$\pm$.005 | No clear effect |
| Train from scratch on the new features rather than reuse a recovered old-feature model | NOGD · Synthetic datasets | accuracy | Hurt |
| Keep updating the recovered-feature model rather than leave it fixed | ROGD-u · Synthetic datasets | accuracy | Helped |
| Keep the recovered-feature model fixed instead of updating it | ROGD-f · Reuter datasets | accuracy | Hurt |
| Use dynamic selection rather than weighted combination | FESL-s vs FESL-c · Average cumulative loss, reported datasets | average cumulative loss | Mixed |
| Add recovered old-feature predictions to new-feature predictions through an adaptive ensemble | FESL · Synthetic datasets | accuracy | Helped |
| Add recovered old-feature predictions to new-feature predictions through an adaptive ensemble | FESL · Reuter datasets | accuracy | Helped |
Closest papers by meaning
- One-Pass Feature Evolvable Learning with Theoretical Guarantees (ICML 2025 · similarity 0.86)
- Learning from Snapshots of Discrete and Continuous Data Streams (NeurIPS 2024 · similarity 0.83)
- SuBiTO: Synopsis-based Training Optimization for Continuous Real-Time Neural Learning over Big Streaming Data (AAAI 2025 · similarity 0.82)
- Test-time Adaptation in Non-stationary Environments via Adaptive Representation Alignment (NeurIPS 2024 · similarity 0.81)
- Online Boosting Adaptive Learning under Concept Drift for Multistream Classification (AAAI 2024 · similarity 0.81)
- Fast Co-Training under Weak Dependence via Stream-Based Active Learning (ICML 2024 · similarity 0.81)
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.