Research map / Learning theory and algorithms / Prediction and decision losses
Active Learning from Peers
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 |
|---|---|---|---|
| query related peer tasks (PEERsum) in addition to the oracle, vs Independent selective sampling | PEER* · Landmine | ACC: 0.9040 -> 0.9403 | Helped |
| PEERsum label queries vs Independent on Landmine | PEER* · Landmine | #Queries: 1802.8 -> 265.6 | Helped |
| PEERsum vs Independent | PEER* · Sentiment | ACC: 0.7522 -> 0.8141 | Helped |
| PEERsum vs Independent | PEER* · Sentiment | #Queries: 2137.6 -> 1494.4 | Helped |
| PEERsum vs Independent | PEER* · Spam | ACC: 0.8309 -> 0.8497 | Helped |
| PEERone (single most related peer) instead of PEERsum (weighted sum of all peers) (ablation) | PEER* · Landmine/Spam/Sentiment | ACC: 0.9403 -> 0.9377 (Landmine); 0.8497 -> 0.8344 (Spam); 0.8141 -> 0.8120 (Sentiment) | Mixed |
| PEERone vs PEERsum training time (ablation) | PEER* · Landmine | Time (s): 0.38 -> 1.01 | Hurt |
| peer queried by highest confidence instead of most related task (ablation) | PEER* · Multiple datasets | accuracy | Hurt |
| increase b2 to infinity (ablation) | PEER* · PEER* | Sentiment ACC | Hurt |
| restrict oracle query budget to 10%, 20%, 30% of T | PEER* · PEER* vs baselines | Sentiment | Helped |
| PEER* vs SHAMPO with small kappa in parallel setting | PEER* · PEER* | Sentiment test ACC | Helped |
| Random query selection vs Independent selective sampling (ablation) | PEER* · baseline | Sentiment ACC: 0.7522 -> 0.7443 | Hurt |
Closest papers by meaning
- Confidence Weighted Multitask Learning (AAAI 2019 · similarity 0.88)
- Multitask Learning with No Regret: from Improved Confidence Bounds to Active Learning (NeurIPS 2023 · similarity 0.88)
- Online Multitask Learning with Long-Term Memory (NeurIPS 2020 · similarity 0.86)
- When More Data Doesn't Help: Limits of Adaptation in Multitask Learning (ICML 2026 · similarity 0.85)
- Peer Learning: Learning Complex Policies in Groups from Scratch via Action Recommendations (AAAI 2024 · similarity 0.83)
- A Two-Stage Learning-to-Defer Approach for Multi-Task Learning (ICML 2025 · similarity 0.82)
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.