arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2501.16952cs.CLcs.AIcs.LG
收藏

多抽象层级的检索增强生成

Multiple Abstraction Level Retrieve Augment Generation

Brandeis University(布兰迪斯大学)

Zheng Zheng, Xinyi Ni, Pengyu Hong

AI总结:

针对现有RAG方法在单一抽象层级检索导致难以跨层级生成答案的问题,提出多抽象层级(MAL)检索方法,在糖科学领域将问答正确率提升25.739%。

摘要:

由大语言模型(LLM)驱动的检索增强生成(RAG)模型,为适应新数据和知识提供了更快、更具成本效益的解决方案。与预训练LLM相比,它还能提供更专业的响应。然而,大多数现有方法依赖于检索前缀大小的块作为参考来支持问答(Q/A)。这种方法通常被部署在单一抽象层级上满足信息需求,因为它难以在多个抽象层级上生成答案。在RAG设置中,虽然LLM在提供足够细节时能够有效地总结和回答问题,但检索过多信息往往会导致“迷失在中间”问题,并超出token限制。我们提出了一种新颖的RAG方法,该方法使用多抽象层级的块(MAL),包括多句级、段落级、章节级和文档级。我们方法的有效性在未被充分探索的糖科学领域中得到了证明。与传统的单层级RAG方法相比,我们的方法在糖相关论文上将AI评估的问答答案正确率提高了25.739%。

英文摘要:

A Retrieval-Augmented Generation (RAG) model powered by a large language model (LLM) provides a faster and more cost-effective solution for adapting to new data and knowledge. It also delivers more specialized responses compared to pre-trained LLMs. However, most existing approaches rely on retrieving prefix-sized chunks as references to support question-answering (Q/A). This approach is often deployed to address information needs at a single level of abstraction, as it struggles to generate answers across multiple levels of abstraction. In an RAG setting, while LLMs can summarize and answer questions effectively when provided with sufficient details, retrieving excessive information often leads to the 'lost in the middle' problem and exceeds token limitations. We propose a novel RAG approach that uses chunks of multiple abstraction levels (MAL), including multi-sentence-level, paragraph-level, section-level, and document-level. The effectiveness of our approach is demonstrated in an under-explored scientific domain of Glycoscience. Compared to traditional single-level RAG approaches, our approach improves AI evaluated answer correctness of Q/A by 25.739\% on Glyco-related papers.

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。

↑