Research map / Learning theory and algorithms / Combinatorial and search optimization
Rethinking Branching on Exact Combinatorial Optimization Solver: The First Deep Symbolic Discovery Framework
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's abstract only. Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Proposes a deep symbolic discovery framework that automatically discovers interpretable, lightweight branching policies for exact combinatorial optimization solvers under branch and bound.
Problem
ML-based branching for NP-hard combinatorial optimization in branch and bound is costly to train and run and hard to interpret, while human-designed branching policies are cheap but hand-crafted and suboptimal.
Approach
Uses deep symbolic discovery to search for compact symbolic branching rules that retain the efficiency and interpretability of human-designed heuristics while approaching learned-policy performance.
Experimental setup
Baselines: human-designed branching heuristics learning-based branching policies
Builds on: branch and bound strong branching
Closest papers by meaning
- Towards General Algorithm Discovery for Combinatorial Optimization: Learning Symbolic Branching Policy from Bipartite Graph (ICML 2024 · similarity 0.95)
- SMILE: Bridging Continuous Optimization and Discrete Symbolic Recovery (NeurIPS 2026 · similarity 0.87)
- SYMBOL: Generating Flexible Black-Box Optimizers through Symbolic Equation Learning (ICLR 2024 · similarity 0.87)
- Improving Monte Carlo Tree Search for Symbolic Regression (NeurIPS 2025 · similarity 0.86)
- Geometric Algorithms for Neural Combinatorial Optimization with Constraints (NeurIPS 2025 · similarity 0.86)
- $\texttt{STRCMP}$: Integrating Graph Structural Priors with Language Models for Combinatorial Optimization (NeurIPS 2025 · similarity 0.86)
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.