Research map / Large language models / Retrieval and knowledge
Backjump-on-Graph: Empowering Large Language Models with Reinforced Retrospective Exploration for Agentic Knowledge Graph Reasoning
Yunqi Zhang, Shiqi Yan, Zhenzhao Yuan, Wenrui Liang, Yangming Liu, Zhixiao Qi, Tianyi Zhang, Shijie Zhang, Wei-Qiang Zhang, Yongfeng Huang, Haixin Duan, Shuai Chen and 1 more
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 RL stage with hybrid reward after SFT (ablation) | BoG, Llama-3.1-8B · CWQ | Hits@1: 79.4 -> 84.0 | Improved |
| add RL stage with hybrid reward after SFT (ablation) | BoG, Llama-3.1-8B · WebQSP | Hits@1: 88.5 -> 92.7 | Improved |
| add step-wise reward (format + faithfulness) (ablation) | BoG, Llama-3.1-8B · CWQ | Hits@1: 82.7 -> 84.0 | Improved |
| add step-wise reward (ablation) | BoG, Llama-3.1-8B · WebQSP | F1: 79.7 -> 81.3 | Improved |
| add efficiency reward (length penalty) (ablation) | BoG, Llama-3.1-8B · CWQ | F1: 77.4 -> 78.3 | Improved |
| add efficiency reward (ablation) | BoG, Llama-3.1-8B · WebQSP | Hits@1: 91.9 -> 92.7 | Improved |
| apply RL to the SFT agent (ablation) | BoG-SFT · CWQ | average step length: 5.3 -> 3.6 | Improved |
| apply RL to the SFT agent (ablation) | BoG-SFT · CWQ | backjumping counts: 227 -> 302 | Improved |
| apply RL to the SFT agent (ablation) | BoG-SFT · CWQ | faithfulness: 0.78 -> 0.89 | Improved |
| fine-tune on oracle-conditioned synthetic backjump trajectories (BoG-SFT) vs closed-source LLMs in the same framework | Gemini-2.5 Flash / Pro, GPT-5.2 · CWQ | Hits@1: 58.4 -> 79.4 | Improved |
| use BoG instead of agent-based KBQA-o1 | KBQA-o1 Llama-3.1-8B · CWQ | Hits@1: 43.2 -> 84.0 | Improved |
| train with RL on one dataset and test on others (OOD) | BoG vs BoG-SFT · WebQSP/CWQ/GrailQA OOD | Hits@1 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- KBQA-o1 — KBQA-o1: Agentic Knowledge Base Question Answering with Monte Carlo Tree Search (ICML 2025) · better on 2 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
- Debate on Graph: A Flexible and Reliable Reasoning Framework for Large Language Models (AAAI 2025 · similarity 0.91)
- Think-on-Graph: Deep and Responsible Reasoning of Large Language Model on Knowledge Graph (ICLR 2024 · similarity 0.91)
- Reasoning on Graphs: Faithful and Interpretable Large Language Model Reasoning (ICLR 2024 · similarity 0.90)
- Graph-constrained Reasoning: Faithful Reasoning on Knowledge Graphs with Large Language Models (ICML 2025 · similarity 0.90)
- Reasoning of Large Language Models over Knowledge Graphs with Super-Relations (ICLR 2025 · similarity 0.90)
- Plan-on-Graph: Self-Correcting Adaptive Planning of Large Language Model on Knowledge Graphs (NeurIPS 2024 · similarity 0.90)
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.
