arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2310.01558cs.CLcs.AI
收藏

使检索增强语言模型对无关上下文具有鲁棒性

Making Retrieval-Augmented Language Models Robust to Irrelevant Context

Tel Aviv University(特拉维夫大学);Allen Institute for AI(艾伦人工智能研究所)

Ori Yoran, Tomer Wolfson, Ori Ram, Jonathan Berant

AI总结:

本文分析检索增强语言模型在无关上下文下性能下降的问题,并提出基于NLI过滤和自动生成混合上下文数据进行微调的方法,仅需1000个样本即可使模型对无关上下文鲁棒且保持高性能。

摘要:

检索增强语言模型(RALMs)有望构建事实准确、高效且实时更新的语言理解系统。RALMs 的一个重要目标是:当检索到的信息相关时能够提升模型性能,而当信息不相关时不会损害性能。这在多跳推理场景中尤为重要,因为误用无关证据可能导致级联错误。然而,近期研究表明,检索增强有时会对性能产生负面影响。本文对五个开放域问答基准进行了全面分析,刻画了检索导致准确率下降的情形。随后,我们提出两种方法来缓解该问题。第一种是一个简单基线,根据自然语言推理(NLI)模型过滤掉不支持问答对的检索段落。该方法能有效防止性能下降,但代价是同时丢弃了相关段落。因此,我们提出一种自动生成数据的方法,通过在训练时混合相关与无关上下文,对语言模型进行微调,使其能够恰当地利用检索到的段落。实验表明,仅需 1,000 个样本就足以训练模型对无关上下文具有鲁棒性,同时在包含相关上下文的样本上保持高性能。

英文摘要:

Retrieval-augmented language models (RALMs) hold promise to produce language understanding systems that are are factual, efficient, and up-to-date. An important desideratum of RALMs, is that retrieved information helps model performance when it is relevant, and does not harm performance when it is not. This is particularly important in multi-hop reasoning scenarios, where misuse of irrelevant evidence can lead to cascading errors. However, recent work has shown that retrieval augmentation can sometimes have a negative effect on performance. In this work, we present a thorough analysis on five open-domain question answering benchmarks, characterizing cases when retrieval reduces accuracy. We then propose two methods to mitigate this issue. First, a simple baseline that filters out retrieved passages that do not entail question-answer pairs according to a natural language inference (NLI) model. This is effective in preventing performance reduction, but at a cost of also discarding relevant passages. Thus, we propose a method for automatically generating data to fine-tune the language model to properly leverage retrieved passages, using a mix of relevant and irrelevant contexts at training time. We empirically show that even 1,000 examples suffice to train the model to be robust to irrelevant contexts while maintaining high performance on examples with relevant ones.

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。

↑