Research map / Learning theory and algorithms / Prediction and decision losses
A Risk Decomposition Framework for Pre-hoc Fine-tuning Prediction
Yuxiang Luo, Chen Wang, Nan Tang
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 |
|---|---|---|---|
| probe fine-tuning for more optimization steps on Dynamic-Critical tasks such as GSM8K | pre-hoc fine-tuning predictors · GSM8K-like reasoning fine-tuning tasks | prediction uncertainty | Improved |
| probe deeper on Static-Sufficient tasks such as SST-2 | pre-hoc fine-tuning predictors · SST-2 and GLUE-style classification | prediction uncertainty | No clear effect |
| probe deeper on Noise-Dominant tasks with label noise | pre-hoc fine-tuning predictors · noisy or ambiguous supervision tasks | prediction uncertainty | No clear effect |
| increase probing depth in general (ablation) | budget-optimal probing · all three regimes | normalized marginal gain of probing | Mixed |
| use shallow fixed-step probing on a Dynamic-Critical task | probing predictors · GSM8K-like tasks | prediction stability | Hurt |
Closest papers by meaning
- Maximizing the Value of Predictions in Control: Accuracy Is Not Enough (NeurIPS 2025 · similarity 0.85)
- DFF: Decision-Focused Fine-Tuning for Smarter Predict-Then-Optimize with Limited Data (AAAI 2025 · similarity 0.84)
- Tuning-Free Stochastic Optimization (ICML 2024 · similarity 0.84)
- Cost-Sensitive Freeze-thaw Bayesian Optimization for Efficient Hyperparameter Tuning (NeurIPS 2025 · similarity 0.84)
- Parametric Scaling Law of Tuning Bias in Conformal Prediction (ICML 2025 · similarity 0.84)
- Tight Lower Bounds and Improved Convergence in Performative Prediction (NeurIPS 2025 · 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.
