paper-with-me

홈 › Papers

Grounding LLM Reasoning under Incomplete Graph Evidence

2026-06-29 · Jiaqi Li, Fanghui Song arxiv

Knowledge graphs can guide large language models (LLMs) reasoning, but the graph seen by a system is usually a retrieved, linked, temporally scoped, and incomplete evidence state rather than a complete account of truth. We develop a theoretical perspective on grounding observable LLM trajectories under such incomplete graph evidence.The evidence state induces entity anchors, typed relation residuals, path energies, and support regions, while the language model supplies a prior over candidate trajectories. We show that, under open-world incompleteness, no hard rule based only on the observed state can both reject every false unsupported trajectory and retain every true-but-unobserved one.We then characterize soft grounding as a KL-regularized deformation of the LLM prior: finite slack preserves support for unsupported but non-contradicted trajectories, whereas hard conditioning appears as an infinite-penalty limit.The framework also yields stability bounds under evidence perturbations and clarifies the constraint regimes appropriate for GraphRAG, KGQA, graph agents, constrained decoding, and faithful generation. The claims are evidence-relative: KG compatibility is treated as declared support, not factual truth.

📄 PDF Abstract BibTeX arXiv:2606.30247

Code (0)

등록된 구현이 없습니다.

Tasks

Knowledge Graphs

Similar Papers 제목 키워드 기반

EFlow: Learning Evidence Flow for Long-Video Reasoning with Adaptive Reflection

2026-07-01 · Wenhao Zhang, Kuanwei Lin, Xuyi Yang, Wei Gao 외 arxiv

Long-video reasoning is fundamentally constrained by how models acquire and utilize visual evidence. Existing tool-augmented video frameworks often interleave temporal grounding and answer reasoning within a single traje…

Reinforcement LearningLogical Reasoning

Asking like Socrates: Socrates helps VLMs understand remote sensing images

2025-11-27 · Run Shao, Ziyu Li, Zhaoyang Zhang, Linrui Xu 외 arxiv

Recent multimodal reasoning models, inspired by DeepSeek-R1, have significantly advanced vision-language systems. However, in remote sensing (RS) tasks, we observe widespread pseudo reasoning: models narrate the process …

Multimodal Reasoning

Relink: Constructing Query-Driven Evidence Graph On-the-Fly for GraphRAG

2026-01-12 · Manzong Huang, Chenyang Bu, Yi He, Xingrui Zhuo 외 arxiv

Graph-based Retrieval-Augmented Generation (GraphRAG) mitigates hallucinations in Large Language Models (LLMs) by grounding them in structured knowledge. However, current GraphRAG methods are constrained by a prevailing …

Open-Domain Question Answering

Topology-Aware Reasoning over Incomplete Knowledge Graph with Graph-Based Soft Prompting

2026-04-14 · Shuai Wang, Xixi Wang, Yinan Yu arxiv

Large Language Models (LLMs) have shown remarkable capabilities across various tasks but remain prone to hallucinations in knowledge-intensive scenarios. Knowledge Base Question Answering (KBQA) mitigates this by groundi…

Knowledge Base Question AnsweringGraph Neural NetworkAnswer GenerationKnowledge Graphs

REFACT: Adaptive Fact Restatement for Compact and Faithful Chain-of-Thought Reasoning

2026-07-23 · Zhensheng Jin, Xin Dai, Zhenghao Liu, Chaojun Xiao 외 arxiv

Large Language Models (LLMs) increasingly leverage long-form reasoning to solve complex tasks, yet their reasoning processes can deviate from the provided context when evidence is incomplete, noisy, or conflicts with par…

Question Answering