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 2404.02581cs.CLcs.IR
收藏

多粒度引导的解码器内融合(Multi-Granularity Guided Fusion-in-Decoder)

Multi-Granularity Guided Fusion-in-Decoder

Sungkyunkwan University(成均馆大学)

Eunseong Choi, Hyeri Lee, Jongwuk Lee

AI总结:

针对开放域问答中 Fusion-in-Decoder 易受看似合理但无关上下文误导的问题,提出 MGFiD,通过段落重排序与句子分类的多任务学习实现多粒度证据识别,并复用重排序结果剪枝段落,在 NQ 和 TQA 上取得优于现有模型的性能。

摘要:

在开放域问答(Open-domain Question Answering, ODQA)中,必须从检索结果中识别出作为证据的相关上下文,并避免虚假上下文。在解码阶段使用拼接的多个上下文的模型架构(即 Fusion-in-Decoder)表现出良好性能,但会因看似合理的上下文而生成错误输出。为解决该问题,我们提出多粒度引导的 Fusion-in-Decoder(MGFiD),在多个粒度层级上识别证据。基于多任务学习,MGFiD 协调段落重排序与句子分类。它将证据性句子聚合为锚点向量,用于指导解码器。此外,它通过复用段落重排序结果进行段落剪枝,提高解码效率。实验表明,MGFiD 在 Natural Questions(NQ)和 TriviaQA(TQA)数据集上优于现有模型,凸显了其多粒度方案的优势。

英文摘要:

In Open-domain Question Answering (ODQA), it is essential to discern relevant contexts as evidence and avoid spurious ones among retrieved results. The model architecture that uses concatenated multiple contexts in the decoding phase, i.e., Fusion-in-Decoder, demonstrates promising performance but generates incorrect outputs from seemingly plausible contexts. To address this problem, we propose the Multi-Granularity guided Fusion-in-Decoder (MGFiD), discerning evidence across multiple levels of granularity. Based on multi-task learning, MGFiD harmonizes passage re-ranking with sentence classification. It aggregates evident sentences into an anchor vector that instructs the decoder. Additionally, it improves decoding efficiency by reusing the results of passage re-ranking for passage pruning. Through our experiments, MGFiD outperforms existing models on the Natural Questions (NQ) and TriviaQA (TQA) datasets, highlighting the benefits of its multi-granularity solution.

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。

↑