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

TL;DR

PathHD introduces an encoder-free, hyperdimensional computing approach for knowledge-graph reasoning with LLMs, significantly reducing latency and GPU usage while maintaining or improving accuracy and interpretability.

Contribution

It proposes a novel encoder-free framework combining hyperdimensional path representations with a single LLM call for efficient, interpretable KG reasoning.

Findings

Matches or improves Hits@1 on WebQSP, CWQ, GrailQA

Reduces end-to-end latency by 40-60%

Lowers GPU memory usage by 3-5 times

Abstract

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.