Research map / Learning theory and algorithms / Prediction and decision losses
Robust Decision-Making with Partially Calibrated Forecasters
Shayan Kiyani, Hamed Hassani, George J. Pappas, Aaron Roth
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 |
|---|---|---|---|
| use the minimax robust policy instead of plug-in best response under the plug-in-tuned adversary | MLP forecaster, squared-loss trained · Bike Sharing (UCI) | mean utility: 0.393 -> 0.412 | Improved |
| use the robust policy instead of plug-in best response under the plug-in-tuned adversary | MLP forecaster, squared-loss trained · California Housing | mean utility: 0.155 -> 0.166 | Improved |
| use the robust policy under the robust-tuned adversary | MLP forecaster, squared-loss trained · Bike Sharing (UCI) | mean utility: 0.402 -> 0.410 | Improved |
| use the robust policy under the robust-tuned adversary | MLP forecaster, squared-loss trained · California Housing | mean utility: 0.160 -> 0.164 | Improved |
| use the robust policy instead of plug-in best response under i.i.d. evaluation | MLP forecaster, squared-loss trained · Bike Sharing (UCI) | mean utility: 0.474 -> 0.463 | Hurt |
| use the robust policy instead of plug-in best response under i.i.d. evaluation | MLP forecaster, squared-loss trained · California Housing | mean utility: 0.216 -> 0.207 | Hurt |
| require decision calibration (or any stronger calibration) of the forecaster | robust decision rule · theory | optimality of plug-in best response | Improved |
Closest papers by meaning
- Truthfulness of Calibration Measures (NeurIPS 2024 · similarity 0.88)
- Algorithms with Calibrated Machine Learning Predictions (ICML 2025 · similarity 0.86)
- Optimal Multiclass U-Calibration Error and Beyond (NeurIPS 2024 · similarity 0.86)
- Testing Calibration in Nearly-Linear Time (NeurIPS 2024 · similarity 0.86)
- High-Dimensional Calibration from Swap Regret (NeurIPS 2025 · similarity 0.85)
- Performative Risk Control: Calibrating Models for Reliable Deployment under Performativity (NeurIPS 2025 · similarity 0.85)
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.
