arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2403.08484cs.CL
收藏

定向高效微调:基于数据驱动样本选择的参数更新优化

Targeted Efficient Fine-tuning: Optimizing Parameter Updates with Data-Driven Sample Selection

Hubei Provincial Key Laboratory of Artificial Intelligence and Smart Learning(湖北省人工智能与智能学习重点实验室);School of Computer, Central China Normal University(华中师范大学计算机学院);National Language Resources Monitoring and Research Center for Network Media(国家网络媒体语言资源监测与研究中心);School of Computer Science and Technology, Huazhong University of Science and Technology(华中科技大学计算机科学与技术学院)

Ming Dong, Kang Xue, Bolong Zheng, Tingting He

AI总结:

针对参数高效微调中随机数据样本选择无法适应不稳定数据分布的问题,提出迭代范围递减(IRD)算法,通过识别高Fisher信息的样本和参数子集优化FISH Mask的样本-参数对选择,在GLUE基准上取得更优性能。

摘要:

对大型语言模型(LLMs)的所有参数进行微调在计算上代价高昂。参数高效微调(PEFT)方法通过选择性地微调特定参数来解决这一问题。大多数参数高效微调(PEFT)方法集中于选择或引入一组待微调的参数。然而,考虑数据样本对参数选择影响的方法很少。具有代表性的数据驱动方法包括基于 FISH Mask 的方法,该方法在选择参数时随机选择一部分数据样本作为基础。然而,这种随机数据样本选择方法无法为不稳定的数据分布选择最优参数。在本研究中,我们引入了一种以数据为中心的方法,并提出了迭代范围递减(IRD)算法来优化 FISH Mask 中的样本-参数对选择。IRD 通过识别具有更高 Fisher 信息的样本和参数子集,迭代地细化选择。我们通过在 GLUE 基准上进行实验,证明了所提策略的有效性和合理性。实验结果表明,我们的策略优化了参数选择,并在一些典型的基线方法上取得了更优的性能。

英文摘要:

Fine-tuning all parameters of Large Language Models (LLMs) is computationally expensive. Parameter-Efficient Fine-Tuning (PEFT) methods address this by selectively fine-tuning specific parameters. Most of the parameter efficient fine-tuning (PEFT) methods center on selecting or introducing a set of parameters to be fine-tuned. However, there are few methods that consider the impact of data samples on parameter selecting. Representative data driven methods include FISH Mask based method, which randomly selects a portion of data samples as a basis when selecting parameters. However, this random data sample selection method cannot select optimal parameters for unstable data distribution. In this work, we introduce a data-centric approach and propose the Iterative Range Decreasing (IRD) algorithm to optimize the sample-parameter pair selection in FISH Mask. IRD iteratively refines the selection by identifying subsets of samples and parameters exhibiting higher Fisher information. We demonstrate the effectiveness and rationality of proposed strategy by conducting experiments on GLUE benchmark. Experimental results show our strategy optimizes the parameter selection and achieves preferable performance over some typical baseline methods.

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。

↑