This paper studies multi-task training of retrieval-augmented generation models for knowledge-intensive tasks. We propose to clean the training set by utilizing a distinct property of knowledge-intensive generation: The connection of query-answer pairs to items in the knowledge base. We filter training examples via a threshold of confidence on the relevance labels, whether a pair is answerable by the knowledge base or not. We train a single Fusion-in-Decoder (FiD) generator on seven combined tasks of the KILT benchmark. The experimental results suggest that our simple yet effective approach substantially improves competitive baselines on two strongly imbalanced tasks; and shows either smaller improvements or no significant regression on the remaining tasks. Furthermore, we demonstrate our multi-task training with relevance label sampling scales well with increased model capacity and achieves state-of-the-art results in five out of seven KILT tasks.

该研究研究了面向知识密集型任务的检索增强式生成模型的多任务训练。通过利用知识密集型生成的独特属性：查询-答案对与知识库中项目的联系，我们提出了一种清理训练集的方法。使用关联性标签的置信度阈值过滤训练样本，以确定是否可以通过知识库回答一对查询-答案。我们在KILT基准测试的七个组合任务上对单个FiD生成器进行训练。实验结果表明，我们简单而有效的方法大大改进了两个强烈不平衡的任务上的竞争基线，并在其余任务上显示出较小的改进或无显着退步。此外，我们表明，与增加的模型容量一样，我们的关联标签采样的多任务训练能够良好地缩放，并在KILT任务的五项中取得了最佳结果。

使用相关采样的多任务检索增强文本生成