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 2104.08691cs.CL
收藏

刻度在参数高效提示微调中的作用

The Power of Scale for Parameter-Efficient Prompt Tuning

Google Research(谷歌研究院)

Brian Lester, Rami Al-Rfou, Noah Constant

摘要:

在这项工作中,我们探索了“提示微调”,这是一种简单而有效的机制,用于学习“软提示”来调节冻结的语言模型以执行特定的下游任务。与GPT-3使用的离散文本提示不同,软提示是通过反向传播学习的,可以调整以包含来自任何数量的标记示例的信号。我们的端到端学习方法远远优于GPT-3的“Few-Shot”学习。更值得注意的是,通过使用T5对模型大小进行烧蚀,我们表明提示微调变得更具规模竞争力:当模型超过数十亿个参数时,我们的方法“缩小了差距”,并与模型调优(其中所有模型权重都被调优)的强大性能相匹配。这一发现尤其重要,因为大型模型的共享和服务成本很高,而将一个冻结模型重用于多个下游任务的能力可以减轻这一负担。我们的方法可以被看作是Li和Liang(2021)最近提出的“前缀调整”的简化,我们提供了与这种方法和其他类似方法的比较。最后,我们表明,空调冻结模型与软提示赋予域转移的鲁棒性的好处,相比,完整的模型调整。

英文摘要:

In this work, we explore "prompt tuning", a simple yet effective mechanism for learning "soft prompts" to condition frozen language models to perform specific downstream tasks. Unlike the discrete text prompts used by GPT-3, soft prompts are learned through backpropagation and can be tuned to incorporate signal from any number of labeled examples. Our end-to-end learned approach outperforms GPT-3's "few-shot" learning by a large margin. More remarkably, through ablations on model size using T5, we show that prompt tuning becomes more competitive with scale: as models exceed billions of parameters, our method "closes the gap" and matches the strong performance of model tuning (where all model weights are tuned). This finding is especially relevant in that large models are costly to share and serve, and the ability to reuse one frozen model for multiple downstream tasks can ease this burden. Our method can be seen as a simplification of the recently proposed "prefix tuning" of Li and Liang (2021), and we provide a comparison to this and other similar approaches. Finally, we show that conditioning a frozen model with soft prompts confers benefits in robustness to domain transfer, as compared to full model tuning.

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。

↑