Research map / Large language models / Retrieval and knowledge
Knowledge Reasoning Language Model: Unifying Knowledge and Language for Inductive Knowledge Graph Reasoning
Xingrui Zhuo, Jiapu Wang, Gongqing Wu, Zhongyuan Wang, Jichen Zhang, Shirui Pan, Xindong Wu
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 |
|---|---|---|---|
| add knowledge encoder (GNN structural representations) (ablation) | KRLM · FB-V1 (E2E) | Hit@10: 0.614 -> 0.705 | Improved |
| add knowledge decoder (structure-constrained projection head) (ablation) | KRLM · FB-V1 (E2E) | Hit@10: 0.674 -> 0.705 | Improved |
| add knowledge memory in KRL attention (ablation) | KRLM · FB-V1 (E2E) | Hit@10: 0.691 -> 0.705 | Improved |
| use PAA instead of mean pooling for word-level embeddings (ablation) | KRLM · FB-V1 (E2E) | Hit@10: 0.692 -> 0.705 | Improved |
| add KRL distillation loss (ablation) | KRLM · NL-0 (E2E) | Hit@10: 0.554 -> 0.591 | Improved |
| add KL divergence term in loss (ablation) | KRLM · WN-V1 (E2E) | Hit@10: 0.798 -> 0.801 | No clear effect |
| add KL term in loss (other datasets) (ablation) | KRLM · NL-0 (E2E) | Hit@10: 0.533 -> 0.591 | Improved |
| fine-tune pretrained KRLM on target dataset vs zero-shot | KRLM · IndE (12 datasets avg) | Hit@10: 0.738 -> 0.751 | Improved |
| KRLM vs best prior KGFM/LLM-based baselines | KRLM (FT) · IndER (13 datasets avg) | Hit@10: 0.542 -> 0.556 | Improved |
| KRLM transductive training vs MKGL/CSProm-KG | KRLM (E2E) · FB15k-237 / WN18RR (transductive) | Hit@10: 0.568 vs 0.591 (FB15k-237); 0.659 vs 0.678 (WN18RR) | Hurt |
| increase knowledge memory size K beyond 50 (ablation) | KRLM · hyperparameter analysis | Hit@10 | No clear effect |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 0 of 1 on most shared results.
- MKGL — MKGL: Mastery of a Three-Word Language (NeurIPS 2024) · better on 1 of 2 shared results
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- Graph-constrained Reasoning: Faithful Reasoning on Knowledge Graphs with Large Language Models (ICML 2025 · similarity 0.92)
- Reasoning on Graphs: Faithful and Interpretable Large Language Model Reasoning (ICLR 2024 · similarity 0.91)
- Think-on-Graph: Deep and Responsible Reasoning of Large Language Model on Knowledge Graph (ICLR 2024 · similarity 0.91)
- LightPROF: A Lightweight Reasoning Framework for Large Language Model on Knowledge Graph (AAAI 2025 · similarity 0.90)
- KG-FIT: Knowledge Graph Fine-Tuning Upon Open-World Knowledge (NeurIPS 2024 · similarity 0.90)
- Towards Graph Foundation Models: Training on Knowledge Graphs Enables Transferability to General Graphs (NeurIPS 2025 · similarity 0.89)
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.
