Research map / Learning theory and algorithms / Combinatorial and search optimization
An Integer Linear Programming Framework for Mining Constraints from Data
Tao Meng, Kai-Wei Chang
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 |
|---|---|---|---|
| mine Sudoku equality constraints and apply them at inference | Outer + EQ · Original and permuted Sudoku | entry-level accuracy: 98.3% -> 100% | Improved |
| add equality constraint mining (ablation) | Outer on MST · MST, 10,000 training samples | Test EM: 71.1% -> 72.9% | Improved |
| add equality constraint mining (ablation) | Outer on MST · MST, 20,000 training samples | Test EM: 87.2% -> 91.8% | Improved |
| use the outer polytope rather than the inner polytope (ablation) | MST constraint mining · MST, 10,000 training samples | Test EM: 41.6% -> 71.1% | Improved |
| use the outer polytope rather than the inner polytope (ablation) | MST constraint mining · MST, 20,000 training samples | Test EM: 69.2% -> 87.2% | Improved |
| add mined constraints to the neural baseline | MST prediction · MST, 20,000 training samples | Test EM: 10.4% -> 91.8% | Improved |
| use latent pairwise label variables with outer-polytope constraints (ablation) | Hierarchical multi-label classification · ImCLEF07A | Test EM: 71.3% -> 79.8% | Improved |
| use latent pairwise label variables with outer-polytope constraints (ablation) | Hierarchical multi-label classification · ImCLEF07A | Test Feasibility: 89.7% -> 100% | Improved |
| use outer-polytope constraints | Baseline neural model · ImCLEF07A | Test EM: 62.6% -> 71.3% | Improved |
| use outer-polytope constraints | Baseline neural model · ImCLEF07A | Test Label Acc.: 98.92% -> 98.85% | No clear effect |
| increase MST training samples from 10,000 to 20,000 for Outer + EQ | Outer + EQ · MST | Test EM: 72.9% -> 91.8% | Improved |
Closest papers by meaning
- Learning to Learn in Interactive Constraint Acquisition (AAAI 2024 · similarity 0.86)
- BIPNN: Learning to Solve Binary Integer Programming via Hypergraph Neural Networks (NeurIPS 2025 · similarity 0.85)
- Geometric Algorithms for Neural Combinatorial Optimization with Constraints (NeurIPS 2025 · similarity 0.84)
- What Are the Rules? Discovering Constraints from Data (AAAI 2024 · similarity 0.84)
- SymILO: A Symmetry-Aware Learning Framework for Integer Linear Optimization (NeurIPS 2024 · similarity 0.84)
- Learning Valid Dual Bounds in Constraint Programming: Boosted Lagrangian Decomposition with Self-Supervised Learning (AAAI 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.
