arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2104.08247cs.CL
收藏

什么是预训练?高效的中间任务选择

What to Pre-Train on? Efficient Intermediate Task Selection

Technical University of Darmstadt(达姆施塔特工业大学)

Clifton Poth, Jonas Pfeiffer, Andreas Rücklé, Iryna Gurevych

摘要:

中间任务微调已被证明可以在许多NLP任务中实现大量的转移增益。由于有大量的候选数据集以及预先训练的语言模型,运行所有组合的叉积以找到最佳转换设置已经变得不可行。在这项工作中,我们首先建立了类似的顺序微调增益可以在适配器设置中实现,随后巩固先前提出的方法,有效地识别中间迁移学习的有益任务。我们实验了一组不同的42个中级和11个目标英语分类,多项选择,问答,和序列标记任务。我们的研究结果表明,有效的嵌入方法,仅依赖于各自的数据集优于计算昂贵的Few-Shot微调方法。我们最好的方法在所有目标任务中实现了不到1%的平均Regret@3,这表明我们能够有效地识别用于中间训练的最佳数据集。

英文摘要:

Intermediate task fine-tuning has been shown to culminate in large transfer gains across many NLP tasks. With an abundance of candidate datasets as well as pre-trained language models, it has become infeasible to run the cross-product of all combinations to find the best transfer setting. In this work we first establish that similar sequential fine-tuning gains can be achieved in adapter settings, and subsequently consolidate previously proposed methods that efficiently identify beneficial tasks for intermediate transfer learning. We experiment with a diverse set of 42 intermediate and 11 target English classification, multiple choice, question answering, and sequence tagging tasks. Our results show that efficient embedding based methods that rely solely on the respective datasets outperform computational expensive few-shot fine-tuning approaches. Our best methods achieve an average Regret@3 of less than 1% across all target tasks, demonstrating that we are able to efficiently identify the best datasets for intermediate training.

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。

↑