AtlasKV：在20GB VRAM中通过十亿级知识图谱增强大语言模型
AtlasKV: Augmenting LLMs with Billion-Scale Knowledge Graphs in 20GB VRAM
AI总结：
本文提出AtlasKV，一种通过十亿级知识图谱增强大语言模型的方法，利用子线性时间与内存复杂度实现高效知识整合，无需外部检索模块或长上下文先验。
摘要：
检索增强生成（RAG）在增强大语言模型（LLM）外部知识方面取得了一些成功。然而，作为一种非参数知识整合范式，RAG方法严重依赖外部检索模块和检索文本上下文先验。特别是对于非常大的知识增强，它们会由于昂贵的搜索和更长的相关上下文而引入显著的推理延迟。在本文中，我们提出了一种参数化知识整合方法，称为AtlasKV，这是一种可扩展、有效且通用的方法，通过极小的GPU内存成本（例如少于20GB VRAM）来使用十亿级知识图谱（KGs）（例如1B三元组）来增强LLM。在AtlasKV中，我们引入KG2KV和HiKVP，以子线性时间与内存复杂度将KG三元组整合到LLM中。它利用LLM固有的注意力机制保持强大的知识基础和泛化性能，并且在适应新知识时不需要外部检索器、长上下文先验或重新训练。
英文摘要：
Retrieval-augmented generation (RAG) has shown some success in augmenting large language models (LLMs) with external knowledge. However, as a non-parametric knowledge integration paradigm for LLMs, RAG methods heavily rely on external retrieval modules and the retrieved textual context prior. Especially for very large scale knowledge augmentation, they would introduce substantial inference latency due to expensive searches and much longer relevant context. In this paper, we propose a parametric knowledge integration method, called \textbf{AtlasKV}, a scalable, effective, and general way to augment LLMs with billion-scale knowledge graphs (KGs) (e.g. 1B triples) using very little GPU memory cost (e.g. less than 20GB VRAM). In AtlasKV, we introduce KG2KV and HiKVP to integrate KG triples into LLMs at scale with sub-linear time and memory complexity. It maintains strong knowledge grounding and generalization performance using the LLMs' inherent attention mechanism, and requires no external retrievers, long context priors, or retraining when adapting to new knowledge.
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。
