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 2401.00368cs.CLcs.IR
收藏

利用大型语言模型改进文本嵌入

Improving Text Embeddings with Large Language Models

Microsoft Corporation(微软公司)

Liang Wang, Nan Yang, Xiaolong Huang, Linjun Yang, Rangan Majumder, Furu Wei

AI总结:

本文提出一种仅用合成数据和不到1k步训练获取高质量文本嵌入的方法,利用专有LLMs生成多语言合成数据微调开源模型,在BEIR和MTEB等基准上取得SOTA结果。

摘要:

在本文中,我们提出了一种新颖且简单的方法,仅使用合成数据和不到1k个训练步骤即可获得高质量的文本嵌入。现有方法通常依赖数十亿弱监督文本对的多阶段中间预训练,随后使用少量标注数据集进行微调,而我们的方法无需构建复杂的训练流程,也不依赖常受限于任务多样性和语言覆盖范围的人工收集数据集。我们利用专有LLMs为93种语言的数十万个文本嵌入任务生成多样化的合成数据。随后,我们使用标准对比损失在合成数据上对开源的decoder-only LLMs进行微调。实验表明,我们的方法在不使用任何标注数据的情况下,在竞争激烈的文本嵌入基准上取得了强劲的表现。此外,当使用合成数据和标注数据的混合进行微调时,我们的模型在BEIR和MTEB基准上创下了新的SOTA结果。

英文摘要:

In this paper, we introduce a novel and simple method for obtaining high-quality text embeddings using only synthetic data and less than 1k training steps. Unlike existing methods that often depend on multi-stage intermediate pre-training with billions of weakly-supervised text pairs, followed by fine-tuning with a few labeled datasets, our method does not require building complex training pipelines or relying on manually collected datasets that are often constrained by task diversity and language coverage. We leverage proprietary LLMs to generate diverse synthetic data for hundreds of thousands of text embedding tasks across 93 languages. We then fine-tune open-source decoder-only LLMs on the synthetic data using standard contrastive loss. Experiments demonstrate that our method achieves strong performance on highly competitive text embedding benchmarks without using any labeled data. Furthermore, when fine-tuned with a mixture of synthetic and labeled data, our model sets new state-of-the-art results on the BEIR and MTEB benchmarks.

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。

↑