Research map / Learning theory and algorithms / Prediction and decision losses
Multidimensional Shape Constraints
Maya Gupta, Erez Louidor, Oleksandr Mangylov, Nobu Morioka, Taman Narayan, Sen Zhao
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 |
|---|---|---|---|
| add 2D unimodality constraints on humidity-month slices (ablation) | Calibrated lattice on Mount Rainier Climbing · Mount Rainier Climbing (Kaggle) | Test MSE: 0.030 to 0.027 | Improved |
| add range dominance (number of reviews over price, price over rating) (ablation) | Calibrated lattice on Play Store installs, non-IID split · Play Store App Installs (non-IID) | Test MSE: 1.276 to 1.250 (vs Mono.) | Improved |
| add Edgeworth constraint (reviews and rating complements) (ablation) | Calibrated lattice on Play Store installs, non-IID split · Play Store App Installs (non-IID) | Test MSE: 1.276 to 1.279 (vs Mono.) | No clear effect |
| add all shape constraints (ablation) | Calibrated lattice on Play Store installs, non-IID split · Play Store App Installs (non-IID) | Test MSE: 1.279 to 1.252 (vs Unconstrained) | Improved |
| add Edgeworth constraint on top of monotonicity (ablation) | Calibrated lattice on Play Store installs, IID split · Play Store App Installs (IID 80/20) | Test MSE: 1.236 to 1.229 (vs Mono.) | No clear effect |
| add range dominance on top of monotonicity (ablation) | Ensemble of calibrated lattices on User Intent (non-IID) · User Intent Prediction (non-IID) | Test MSE: 0.752 to 0.751 | Improved |
| add Edgeworth constraint on top of monotonicity (ablation) | Ensemble of calibrated lattices on User Intent (non-IID) · User Intent Prediction (non-IID) | Test MSE: 0.752 to 0.751 | Improved |
| add trapezoid constraints on top of monotonicity (ablation) | Ensemble of calibrated lattices on User Intent (non-IID) · User Intent Prediction (non-IID) | Test MSE: 0.752 to 0.764 | Hurt |
| add monotonic-dominance-style constraints on lattice ensembles with increasing calibrators (joint monotonicity / monotonic dominance) | Calibrated lattice · User Intent Prediction | Test MSE | No clear effect |
| impose monotonicity and dominance constraints on past-history features of lattice models | Lattice models on Weekly Sales Transactions · Weekly Sales Transactions | Test MSE (figure, qualitative) | Improved |
Closest papers by meaning
- PenDer: Incorporating Shape Constraints via Penalized Derivatives (AAAI 2021 · similarity 0.86)
- Diminishing Returns Shape Constraints for Interpretability and Regularization (NeurIPS 2018 · similarity 0.85)
- StructEdit: Learning Structural Shape Variations (CVPR 2020 · similarity 0.85)
- Shape Constraints for Set Functions (ICML 2019 · similarity 0.85)
- Enhancing Implicit Shape Generators Using Topological Regularizations (ICML 2024 · similarity 0.83)
- Empowering Decision Trees via Shape Function Branching (NeurIPS 2025 · similarity 0.83)
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.
