Research map / Graph learning / GNN architectures
Unsat Core Prediction through Polarity-Aware Representation Learning over Clause-Literal Hypergraphs
Zhenchao Sun, Shuai Ma, Ping Lu, Chongyang Tao
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 |
|---|---|---|---|
| replace bipartite graph with hypergraph message passing (ablation) | NeuroCore-based paSAT · CA unsat-core | Top-M precision (avg): 0.238 -> 0.360 | Improved |
| add invariant-equivariant variable decomposition (ablation) | paSAT (HG) · CA unsat-core | Top-M precision (avg): 0.360 -> 0.383 | Improved |
| add polarity-inversion consistency regularization (ablation) | paSAT (HG+DE) · CA unsat-core | Top-M precision (avg): 0.383 -> 0.408 | Improved |
| replace bipartite graph with hypergraph message passing (ablation) | NeuroCore-based paSAT · SR unsat-core | Top-M precision (avg): 0.895 -> 0.898 | Improved |
| add decomposition and regularization to hypergraph model (ablation) | paSAT (HG) · SR unsat-core | ROC-AUC (avg): 0.808 -> 0.854 -> 0.872 | Improved |
| replace bipartite graph with hypergraph message passing (ablation) | NeuroCore-based paSAT · PS unsat-core | Top-M precision (avg): 0.774 -> 0.784 | Improved |
| add decomposition and regularization to hypergraph model (ablation) | paSAT (HG) · PS unsat-core | Top-M precision (avg): 0.784 -> 0.792 -> 0.797 | Improved |
| use paSAT instead of NeuroCore | unsat-core predictor · CA unsat-core | Top-M precision (avg): 0.238 -> 0.408 | Improved |
| use paSAT instead of NeuroCore | unsat-core predictor · SR unsat-core | PR-AUC (avg): 0.938 -> 0.963 | Improved |
| use SATFormer instead of GCN | unsat-core predictor · SR unsat-core | Top-M precision (avg): 0.880 -> 0.881 | No clear effect |
| guide Glucose with paSAT variable scores | Glucose · SATCOMP 2023 | instances solved: 133 -> 148 | Improved |
| guide Glucose with paSAT variable scores | Glucose · SATCOMP 2024 | instances solved: 111 -> 118 | Improved |
Closest papers by meaning
- NeuroBack: Improving CDCL SAT Solving using Graph Neural Networks (ICLR 2024 · similarity 0.86)
- Learning Better Representations From Less Data For Propositional Satisfiability (NeurIPS 2024 · similarity 0.86)
- Learning How to Cube (NeurIPS 2026 · similarity 0.86)
- Hardness of Random Reordered Encodings of Parity for Resolution and CDCL (AAAI 2024 · similarity 0.85)
- HardCore Generation: Generating Hard UNSAT Problems for Data Augmentation (NeurIPS 2024 · similarity 0.85)
- The Impact of Literal Sorting on Cardinality Constraint Encodings (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.
