Research map / Learning theory and algorithms / Prediction and decision losses
Hedging on the Frontier: Learning New Tasks with Few Samples
Tobias Wegel, Federico Di Gennaro, Geelon So, Fanny Yang
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 |
|---|---|---|---|
| restrict ERM to the Pareto frontier of source benchmarks (ERM-Pareto) | ERM on a model zoo · HELM Lite (K=3 sources, small target n) | mean excess risk | Improved |
| restrict ERM to the Pareto frontier of source benchmarks (ERM-Pareto) | ERM on a model zoo · VHELM (K=3 sources, small target n) | mean excess risk | Improved |
| use adaptive margin ERM instead of ERM over all models | ERM-all · HELM and VHELM (K=3) | mean excess risk | Improved |
| restrict ERM to the Pareto set (versus all models) at large target sample size (ablation) | ERM-Pareto · HELM and VHELM (K=1, large n) | mean excess risk | Hurt |
| grow the candidate set with n (adaptive margin) instead of fixed Pareto set (ablation) | ERM-Pareto · HELM and VHELM (various K) | mean excess risk | Improved |
| use Min-Max source-only selection instead of using target samples (ablation) | Min-Max baseline · HELM (K=1) | mean excess risk | No clear effect |
| increase the number of source benchmarks K (ablation) | frontier-biased ERM · HELM (K=1..6) and VHELM (K=1..8) | mean excess risk | No clear effect |
Closest papers by meaning
- Scaling Reward Modeling without Human Supervision (NeurIPS 2026 · similarity 0.85)
- The Illusion of Diminishing Returns: Measuring Long Horizon Execution in LLMs (ICLR 2026 · similarity 0.85)
- Ascent Fails to Forget (NeurIPS 2025 · similarity 0.85)
- Weak-to-Strong Generalization Even in Random Feature Networks, Provably (ICML 2025 · similarity 0.85)
- Why Has Predicting Downstream Capabilities of Frontier AI Models with Scale Remained Elusive? (ICML 2025 · similarity 0.85)
- Inherent Trade-Offs between Diversity and Stability in Multi-Task Benchmarks (ICML 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.
