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 2305.18952cs.IRcs.AI
收藏

探索生成式检索在动态语料库上的实用性

Exploring the Practicality of Generative Retrieval on Dynamic Corpora

KAIST AI(韩国科学技术院人工智能学院);LBox(LBox公司);University of Washington(华盛顿大学);UCL(伦敦大学学院);Seoul National University(首尔大学)

Chaeeun Kim, Soyoung Yoon, Hyunji Lee, Joel Jang, Sohee Yang, Minjoon Seo

AI总结:

针对现实中动态语料库的检索需求,该研究探究生成式检索在动态场景的适应性与鲁棒性,经StreamingQA基准测试验证,其相较双编码器在知识适配、效率等方面更具优势,为实用检索系统提供新方向。

摘要:

信息检索(IR)的性能基准测试大多基于固定文档集合(静态语料库)开展。但在现实场景中,情况鲜少如此,待检索的文档会持续更新和新增。本文聚焦于将自回归语言模型应用于信息检索问题的生成式检索(Generative Retrievals,GR),探究其在动态场景下的适应性与鲁棒性。我们还对计算和内存效率进行了广泛评估,这是部署处理海量且不断变化的文档集合的现实世界信息检索系统的关键因素。我们在StreamingQA基准上的实验结果表明,与检索系统中常用的双编码器(Dual Encoders,DE)相比,生成式检索对不断演化的知识适应性更强(提升4-11%),在学习带时间信息的知识方面具备鲁棒性,且在推理浮点运算量(FLOPs,为双编码器的2倍)、索引时间(为双编码器的6倍)和存储占用(为双编码器的4倍)方面更高效。本文凸显了生成式检索未来在动态环境下实用信息检索系统中应用的潜力。

英文摘要:

Benchmarking the performance of information retrieval (IR) is mostly conducted with a fixed set of documents (static corpora). However, in realistic scenarios, this is rarely the case and the documents to be retrieved are constantly updated and added. In this paper, we focus on Generative Retrievals (GR), which apply autoregressive language models to IR problems, and explore their adaptability and robustness in dynamic scenarios. We also conduct an extensive evaluation of computational and memory efficiency, crucial factors for real-world deployment of IR systems handling vast and ever-changing document collections. Our results on the StreamingQA benchmark demonstrate that GR is more adaptable to evolving knowledge (4-11%), robust in learning knowledge with temporal information, and efficient in terms of inference FLOPs (x2), indexing time (x6), and storage footprint (x4) compared to Dual Encoders (DE), which are commonly used in retrieval systems. Our paper highlights the potential of GR for future use in practical IR systems within dynamic environments.

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。

↑