arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2408.01119cs.CL
收藏

任务提示向量:通过多任务软提示迁移实现有效的初始化

Task Prompt Vectors: Effective Initialization through Multi-Task Soft-Prompt Transfer

Faculty of Information Technology, Brno University of Technology(布拉格技术大学信息学院);Kempelen Institute of Intelligent Technologies(智能技术研究所);German Research Center for Artificial Intelligence (DFKI)(德国人工智能研究中心)

Robert Belanec, Simon Ostermann, Ivan Srba, Maria Bielikova

AI总结:

本文提出任务提示向量,通过多任务软提示迁移实现有效初始化,适用于低资源环境,并展示了其在不同语言模型架构中的独立性,提供了一种多任务提示算术的替代方案。

摘要:

提示微调是训练大型语言模型(LLMs)的有效方法。然而,当前基于软提示的方法往往牺牲了多任务模块化,需要为每个新任务重复完整的或部分训练过程。尽管最近的任务向量工作通过在完整模型权重上应用算术运算来实现所需的多任务性能,但针对软提示的类似方法仍缺失。为此,我们引入了任务提示向量,其通过元素级差分计算微调后的软提示权重与其随机初始化之间的差异。在12个NLU数据集上的实验结果表明,任务提示向量可以用于低资源环境,有效初始化相似任务的提示微调。此外,我们展示了任务提示向量在两种不同语言模型架构上的提示微调随机初始化中的独立性。这允许使用来自不同任务的预训练向量进行提示算术。通过这种方式,我们通过多个任务的提示向量算术加法提供了一种与最新基线相竞争的替代方案。

英文摘要:

Prompt tuning is an efficient solution for training large language models (LLMs). However, current soft-prompt-based methods often sacrifice multi-task modularity, requiring the training process to be fully or partially repeated for each newly added task. While recent work on task vectors applied arithmetic operations on full model weights to achieve the desired multi-task performance, a similar approach for soft-prompts is still missing. To this end, we introduce Task Prompt Vectors, created by element-wise difference between weights of tuned soft-prompts and their random initialization. Experimental results on 12 NLU datasets show that task prompt vectors can be used in low-resource settings to effectively initialize prompt tuning on similar tasks. In addition, we show that task prompt vectors are independent of the random initialization of prompt tuning on 2 different language model architectures. This allows prompt arithmetics with the pre-trained vectors from different tasks. In this way, we provide a competitive alternative to state-of-the-art baselines by arithmetic addition of task prompt vectors from multiple tasks.

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。

↑