基于生成式伪标注的无监督多语言稠密检索
Unsupervised Multilingual Dense Retrieval via Generative Pseudo Labeling
AI总结：
针对多语言稠密检索依赖大量配对数据的难题，提出无监督模型UMR，利用多语言大模型的序列似然估计生成伪标签，通过两阶段迭代框架训练，在两个基准数据集上超越有监督基线，且代码数据模型已开源。
摘要：
稠密检索方法已在多语言信息检索（查询与文档可使用不同语言）中展现出可观性能。然而，稠密检索器通常需要大量配对数据，这在多语言场景下带来了更严峻的挑战。本文提出UMR——一种无需任何配对数据训练的无监督多语言稠密检索器。我们的方法利用多语言语言模型的序列似然估计能力，获取用于训练稠密检索器的伪标签。我们提出了一个两阶段框架，可迭代提升多语言稠密检索器的性能。在两个基准数据集上的实验结果表明，UMR优于有监督基线方法，证明了无需配对数据训练多语言检索器的潜力，从而提升了其实用性。我们的源代码、数据与模型已在https://github.com/MiuLab/UMR公开。
英文摘要：
Dense retrieval methods have demonstrated promising performance in multilingual information retrieval, where queries and documents can be in different languages. However, dense retrievers typically require a substantial amount of paired data, which poses even greater challenges in multilingual scenarios. This paper introduces UMR, an Unsupervised Multilingual dense Retriever trained without any paired data. Our approach leverages the sequence likelihood estimation capabilities of multilingual language models to acquire pseudo labels for training dense retrievers. We propose a two-stage framework which iteratively improves the performance of multilingual dense retrievers. Experimental results on two benchmark datasets show that UMR outperforms supervised baselines, showcasing the potential of training multilingual retrievers without paired data, thereby enhancing their practicality. Our source code, data, and models are publicly available at https://github.com/MiuLab/UMR
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。
