arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2310.20127cs.CL
收藏

利用可学习提示层改进提示微调

Improving Prompt Tuning with Learned Prompting Layers

East China Normal University(华东师范大学);Southern University of Science and Technology(南方科技大学)

Wei Zhu, Ming Tan

AI总结:

本文提出选择性提示微调框架SPT,通过可学习概率门自动选择提示层,并结合SPT-DARTS双层优化提升性能,在多个基准上超越现有方法。

摘要:

提示微调将软提示前置到输入嵌入或隐藏状态,并仅优化提示以使预训练模型(PTMs)适应下游任务。以往的工作手动选择提示层,远非最优,未能充分挖掘提示微调的潜力。在本工作中,我们提出了一种新颖的框架,选择性提示微调(SPT),它通过在每个中间层插入一个由可学习概率门控制的提示来学习选择合适的提示层。我们进一步提出了一种新颖的双层优化框架SPT-DARTS,能够更好地优化可学习门,并改善所学提示层设置的最终提示微调性能。我们在全数据和少样本场景下对十个基准数据集进行了广泛实验。结果表明,我们的SPT框架在可调参数相当或更少的情况下,性能优于先前最先进的PETuning基线。

英文摘要:

Prompt tuning prepends a soft prompt to the input embeddings or hidden states and only optimizes the prompt to adapt pretrained models (PTMs) to downstream tasks. The previous work manually selects prompt layers which are far from optimal and failed to exploit the potential of prompt tuning. In this work, we propose a novel framework, \underline{S}elective \underline{P}rompt \underline{T}uning (SPT), that learns to select the proper prompt layers by inserting a prompt controlled by a learnable probabilistic gate at each intermediate layer. We further propose a novel bi-level optimization framework, SPT-DARTS, that can better optimize the learnable gates and improve the final prompt tuning performances of the learned prompt layer settings. We conduct extensive experiments with ten benchmark datasets under the full-data and few-shot scenarios. The results demonstrate that our SPT framework can perform better than the previous state-of-the-art PETuning baselines with comparable or fewer tunable parameters.

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。

↑