Research map / Large language models / Retrieval and knowledge
CLAUSE: Agentic Neuro-Symbolic Knowledge Graph Reasoning via Dynamic Learnable Context Engineering
Yang Zhao, Chengxiao Dai, Wei Zhuo, Yue Xiu, Dusit Niyato
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 Subgraph Architect (ablation) | CLAUSE · MetaQA | EM@1: 74.8 -> 87.3 | Improved |
| add Subgraph Architect (ablation) | CLAUSE · MetaQA | latency (x): 1.32 -> 1.00 | Improved |
| add Subgraph Architect (ablation) | CLAUSE · MetaQA | edge budget (x): 1.44 -> 1.00 | Improved |
| add Path Navigator (ablation) | CLAUSE · MetaQA | EM@1: 82.1 -> 87.3 | Improved |
| add Path Navigator (ablation) | CLAUSE · MetaQA | latency (x): 1.18 -> 1.00 | Improved |
| add Context Curator (ablation) | CLAUSE · MetaQA | EM@1: 80.6 -> 87.3 | Improved |
| add Lagrangian dual updates (LC-MAPPO) over MAPPO (ablation) | CLAUSE · MetaQA | EM@1: 85.0 -> 87.3 | Improved |
| add Lagrangian dual updates (LC-MAPPO) over MAPPO (ablation) | CLAUSE · MetaQA | edge budget (x): 1.28 -> 1.00 | Improved |
| update lambda during training instead of fixing it (ablation) | CLAUSE · MetaQA | EM@1: 84.6 -> 87.3 | Improved |
| use LC-MAPPO instead of MAPPO under constraints | CLAUSE · MetaQA (edge 0.5, latency 0.7) | feasibility rate: 0.117 -> 0.340 | Improved |
| use LC-MAPPO instead of MAPPO under constraints | CLAUSE · MetaQA (edge 0.5, latency 0.7) | latency violations: 0.880 -> 0.577 | Improved |
| use CLAUSE instead of GraphRAG | GraphRAG · MetaQA 2-hop | EM@1: 48.0 -> 87.3 | Improved |
Closest papers by meaning
- Backjump-on-Graph: Empowering Large Language Models with Reinforced Retrospective Exploration for Agentic Knowledge Graph Reasoning (ICML 2026 · similarity 0.91)
- AgentHop: A Diagnostic Benchmark for Agentic Multi-Hop Scientific Question Answering (NeurIPS 2026 · similarity 0.90)
- LEGO: Latent Execution-Guided Reasoning for Multi-Hop Question Answering on Knowledge Graphs (ICML 2021 · similarity 0.89)
- Debate on Graph: A Flexible and Reliable Reasoning Framework for Large Language Models (AAAI 2025 · similarity 0.88)
- Knowledge-Graph Paths as Intermediate Supervision for Self-Evolving Search Agents (NeurIPS 2026 · similarity 0.88)
- Think-on-Graph: Deep and Responsible Reasoning of Large Language Model on Knowledge Graph (ICLR 2024 · similarity 0.87)
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.
