AI Paper Atlas

Research map / Learning theory and algorithms / Prediction and decision losses

Decision-Focused Learning without Decision-Making: Learning Locally Optimized Decision Losses

Sanket Shah, Kai Wang, Bryan Wilder, Andrew Perrault, Milind Tambe

NeurIPS 2022 · Learning theory and algorithms › Prediction and decision losses · approach: Losses for decision-making · 8 citations (OpenAlex)

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 changedWhereMeasuredEffect
Train with DirectedQuadratic LODL instead of 2-stage MSEpredictive model · Web Advertisingnormalized DQ: 0.48 -> 0.91Improved
Train with DirectedQuadratic LODL instead of 2-stage MSEpredictive model · Linear Modelnormalized DQ: -0.95 -> 0.96Improved
Train with DirectedQuadratic LODL instead of 2-stage MSEpredictive model · Portfolio Optimizationnormalized DQ: 0.32 -> 0.33Improved
Replace handcrafted DFL surrogate with learned Quadratic LODLpredictive model · Web Advertisingnormalized DQ: 0.85 -> 0.93Improved
Replace handcrafted DFL surrogate with DirectedQuadratic LODLpredictive model · Linear Modelnormalized DQ: 0.83 -> 0.96Improved
Replace DFL with DirectedQuadratic LODLpredictive model · Portfolio Optimizationnormalized DQ: 0.35 -> 0.33Hurt
Use non-convex NN LODL instead of convex DirectedQuadraticLODL · Portfolio Optimizationnormalized DQ: 0.33 -> -0.11Hurt
Add directionality to WeightedMSE LODLLODL · Linear Modelnormalized DQ: -0.93 -> 0.96Improved
Add directionality to WeightedMSE LODLLODL · Web Advertisingnormalized DQ: 0.58 -> 0.53No clear effect
Add cross-prediction (quadratic) terms to WeightedMSE LODLLODL · Web Advertisingnormalized DQ: 0.58 -> 0.93Improved
Choose sampling strategy 2-Perturbed vs All-Perturbed (ablation)NN and DirectedWeightedMSE LODL · Web Advertisingnormalized DQMixed
Increase number of samples used to fit LODL (ablation)LODL · Web Advertisingnormalized DQImproved

Closest papers by meaning

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.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy