arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2512.09369cs.LGcs.CL
收藏

无需编码器的知识图谱推理:通过超维路径检索实现LLM

Encoder-Free Knowledge-Graph Reasoning with LLMs via Hyperdimensional Path Retrieval

University of California, Irvine(加州大学 Irvine 分校)

Yezi Liu, William Youngwoo Chung, Hanning Chen, Calvin Yeung, Mohsen Imani

AI总结:

PathHD通过超维路径检索实现无需编码器的知识图谱推理,提升效率与可解释性。

摘要:

近年来,大语言模型(LLM)的进步使基于知识的推理变得更加实用,但基于知识图谱(KG)的问答系统往往在效率和透明度上付出高昂代价。在典型流程中,符号路径由神经编码器评分或通过多次LLM调用重新排序,这会增加延迟和GPU成本,并使决策过程难以审计。我们引入PathHD,一种无需编码器的知识图谱推理框架,它将超维计算(HDC)与每个查询一次LLM调用相结合。给定一个查询,PathHD将关系路径表示为块对角线GHRR超向量,使用校准的块级余弦相似度进行候选路径检索,并通过一次LLM裁定输出最终答案以及可引用的路径。该设计得益于三个技术组件:(i)一个顺序敏感、非可交换的绑定运算符,用于组合多跳路径;(ii)一个稳健的相似度校准,稳定超向量检索;(iii)一个裁定阶段,保留可解释性同时避免单路径LLM评分。在WebQSP、CWQ和GrailQA上,PathHD在Hits@1方面与强大的神经基线模型持平或改进,仅使用每个查询一次LLM调用,将端到端延迟降低40-60%,并由于无需编码器检索而降低GPU内存3-5倍。总体而言,结果表明,经过精心设计的HDC路径表示可以作为高效且忠实的KG-LLM推理的有效基质,实现准确率、效率和可解释性的良好权衡。

英文摘要:

Recent progress in large language models (LLMs) has made knowledge-grounded reasoning increasingly practical, yet KG-based QA systems often pay a steep price in efficiency and transparency. In typical pipelines, symbolic paths are scored by neural encoders or repeatedly re-ranked by multiple LLM calls, which inflates latency and GPU cost and makes the decision process hard to audit. We introduce PathHD, an encoder-free framework for knowledge-graph reasoning that couples hyperdimensional computing (HDC) with a single LLM call per query. Given a query, PathHD represents relation paths as block-diagonal GHRR hypervectors, retrieves candidate paths using a calibrated blockwise cosine similarity with Top-K pruning, and then performs a one-shot LLM adjudication that outputs the final answer together with supporting, citeable paths. The design is enabled by three technical components: (i) an order-sensitive, non-commutative binding operator for composing multi-hop paths, (ii) a robust similarity calibration that stabilizes hypervector retrieval, and (iii) an adjudication stage that preserves interpretability while avoiding per-path LLM scoring. Across WebQSP, CWQ, and GrailQA, PathHD matches or improves Hits@1 compared to strong neural baselines while using only one LLM call per query, reduces end-to-end latency by $40-60\%$, and lowers GPU memory by $3-5\times$ due to encoder-free retrieval. Overall, the results suggest that carefully engineered HDC path representations can serve as an effective substrate for efficient and faithful KG-LLM reasoning, achieving a strong accuracy-efficiency-interpretability trade-off.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑