Research map / Probabilistic and scientific modeling / Molecules and proteins
Differentiable Learning of Logical Rules for Knowledge Base Reasoning
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 end-to-end differentiable rule and parameter learning instead of ISG's alternating structure and parameter search | Neural LP vs ISG · UMLS | Hits@10: 43.5 -> 93.2 | Helped |
| Use end-to-end differentiable rule and parameter learning instead of ISG's alternating structure and parameter search | Neural LP vs ISG · Kinship | Hits@10: 59.2 -> 90.2 | Helped |
| Set maximum rule length to 3 rather than 2 (ablation) | Neural LP · UMLS | Hits@10: 92.0 -> 93.2 | Helped |
| Set maximum rule length to 3 rather than 2 (ablation) | Neural LP · Kinship | Hits@10: 90.2 -> 90.1 | Hurt |
| Use Neural LP instead of Node+LinkFeat | Knowledge base completion · WN18 | MRR: 0.94 -> 0.94 | No clear effect |
| Use Neural LP instead of Implicit ReasoNets | Knowledge base completion · WN18 | Hits@10: 95.3 -> 94.5 | Hurt |
| Use Neural LP instead of Node+LinkFeat | Knowledge base completion · FB15K | MRR: 0.82 -> 0.76 | Hurt |
| Use Neural LP instead of Implicit ReasoNets | Knowledge base completion · FB15K | Hits@10: 92.7 -> 83.7 | Hurt |
| Use Neural LP instead of DistMult | Knowledge base completion · FB15KSelected | MRR: 0.25 -> 0.24 | Hurt |
| Use Neural LP instead of DistMult | Knowledge base completion · FB15KSelected | Hits@10: 40.8 -> 36.2 | Hurt |
| Use entity-independent logical rules rather than entity-dependent TransE representations | Neural LP vs TransE · Inductive WN18 | Hits@10: 0.01 -> 94.49 | Helped |
| Use entity-independent logical rules rather than entity-dependent TransE representations | Neural LP vs TransE · Inductive FB15K | Hits@10: 0.48 -> 73.28 | Helped |
Closest papers by meaning
- Differentiable Inductive Logic Programming for Structured Examples (AAAI 2021 · similarity 0.90)
- Differentiable learning of numerical rules in knowledge graphs (ICLR 2020 · similarity 0.89)
- RNNLogic: Learning Logic Rules for Reasoning on Knowledge Graphs (ICLR 2021 · similarity 0.89)
- Differentiable Neuro-Symbolic Reasoning on Large-Scale Knowledge Graphs (NeurIPS 2023 · similarity 0.89)
- An Evaluation of Approaches to Train Embeddings for Logical Inference (Student Abstract) (AAAI 2025 · similarity 0.87)
- Differentiable Rule Induction from Raw Sequence Inputs (ICLR 2025 · similarity 0.85)
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.