Research map / Graph learning / Knowledge and temporal graphs
Embedding Symbolic Knowledge into Deep Networks
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 |
|---|---|---|---|
| use d-DNNF rather than CNF for formula embeddings | LENSR entailment model · Synthetic entailment, moderate | accuracy (%): 71.19 (0.93) -> 82.49 (1.11) | Helped |
| use d-DNNF rather than CNF for formula embeddings | LENSR VRP model with heterogeneous embedding and semantic regularization · VRD | Top-5 accuracy (%): 85.70 -> 92.77 | Helped |
| add semantic regularization (ablation) | LENSR d-DNNF · VRD | Top-5 accuracy (%): 85.37 -> 88.01 | Helped |
| add heterogeneous node embedding (ablation) | LENSR d-DNNF · VRD | Top-5 accuracy (%): 85.37 -> 90.13 | Helped |
| add heterogeneous node embedding and semantic regularization together (ablation) | LENSR d-DNNF · Synthetic entailment, moderate | accuracy (%): 82.49 (1.11) -> 83.04 (1.58) | Helped |
| add semantic regularization (ablation) | LENSR d-DNNF · Synthetic entailment, low | accuracy (%): 89.91 (0.31) -> 90.22 (0.23) | No clear effect |
| add semantic regularization (ablation) | LENSR d-DNNF · Synthetic entailment, moderate | accuracy (%): 82.49 (1.11) -> 82.28 (1.40) | No clear effect |
| add heterogeneous node embedding (ablation) | LENSR d-DNNF · Synthetic entailment, moderate | accuracy (%): 82.49 (1.11) -> 81.30 (1.29) | Hurt |
| add semantic regularization (ablation) | LENSR d-DNNF · Synthetic entailment, high | accuracy (%): 70.56 (1.16) -> 71.46 (1.17) | No clear effect |
| add heterogeneous node embedding (ablation) | LENSR d-DNNF · Synthetic entailment, high | accuracy (%): 70.56 (1.16) -> 70.54 (0.62) | No clear effect |
| add LENSR logic embeddings and logic loss | VRP baseline without logic · VRD | Top-5 accuracy (%): 84.30 -> 92.77 | Helped |
| use direct semantic loss rather than LENSR embeddings | VRP model · VRD | Top-5 accuracy (%): 84.30 -> 84.76 | No clear effect |
Closest papers by meaning
- Symbolic Graph Reasoning Meets Convolutions (NeurIPS 2018 · similarity 0.89)
- Rule-Guided Compositional Representation Learning on Knowledge Graphs (AAAI 2020 · similarity 0.88)
- Towards Foundation Models for Knowledge Graph Reasoning (ICLR 2024 · similarity 0.87)
- Improving Soft Unification with Knowledge Graph Embedding Methods (ICML 2025 · similarity 0.87)
- Systematic Relational Reasoning With Epistemic Graph Neural Networks (ICLR 2025 · similarity 0.86)
- Toward a Unified Statistical Theory of Unsupervised Pretraining and Supervised Neural Knowledge Graph Learning (NeurIPS 2026 · 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.