arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2412.07618cs.AIcs.CL
收藏

适应非平稳环境:知识图谱上多臂老虎机增强的检索增强生成

Adapting to Non-Stationary Environments: Multi-Armed Bandit Enhanced Retrieval-Augmented Generation on Knowledge Graphs

The Hong Kong University of Science and Technology (Guangzhou)(香港科技大学(广州));Tencent Hunyuan(腾讯混元)

Xiaqiang Tang, Jian Li, Nan Du, Sihong Xie

AI总结:

针对非平稳环境中RAG系统性能下降问题,提出多目标多臂老虎机增强的RAG框架,将检索方法作为臂并利用实时反馈动态选择,在平稳与非平稳环境下均取得优异表现。

摘要:

尽管大型语言模型在许多NLP任务上表现出色,但在记忆广泛的世界知识方面仍面临显著局限。最近的研究表明,利用检索增强生成(RAG)框架,结合以结构化格式封装大量事实数据的知识图谱,能够稳健地增强LLM的推理能力。然而,在真实场景中部署此类系统面临挑战:非平稳环境的持续演变可能导致性能下降,且用户满意度需要在性能和响应速度之间进行谨慎平衡。为了应对这些挑战,我们引入了一个多目标多臂老虎机增强的RAG框架,该框架由实践中丰富且不断演变的检索上下文下具有多样能力的多种检索方法提供支持。在该框架中,每种检索方法被视为一个独立的“臂”。系统利用实时用户反馈来适应动态环境,通过根据输入查询和每个臂的历史多目标表现来选择合适的检索方法。在两个基准KGQA数据集上进行的大量实验表明,我们的方法在非平稳设置中显著优于基线方法,同时在平稳环境中实现了最先进的性能。代码和数据可在 https://github.com/FUTUREEEEEE/Dynamic-RAG.git 获取

英文摘要:

Despite the superior performance of Large language models on many NLP tasks, they still face significant limitations in memorizing extensive world knowledge. Recent studies have demonstrated that leveraging the Retrieval-Augmented Generation (RAG) framework, combined with Knowledge Graphs that encapsulate extensive factual data in a structured format, robustly enhances the reasoning capabilities of LLMs. However, deploying such systems in real-world scenarios presents challenges: the continuous evolution of non-stationary environments may lead to performance degradation and user satisfaction requires a careful balance of performance and responsiveness. To address these challenges, we introduce a Multi-objective Multi-Armed Bandit enhanced RAG framework, supported by multiple retrieval methods with diverse capabilities under rich and evolving retrieval contexts in practice. Within this framework, each retrieval method is treated as a distinct ``arm''. The system utilizes real-time user feedback to adapt to dynamic environments, by selecting the appropriate retrieval method based on input queries and the historical multi-objective performance of each arm. Extensive experiments conducted on two benchmark KGQA datasets demonstrate that our method significantly outperforms baseline methods in non-stationary settings while achieving state-of-the-art performance in stationary environments. Code and data are available at https://github.com/FUTUREEEEEE/Dynamic-RAG.git

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。

↑