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

TL;DR

This work introduces a data-centric approach and proposes the Iterative Range Decreasing (IRD) algorithm to optimize the sample-parameter pair selection in FISH Mask and shows its strategy optimizes the parameter selection and achieves preferable performance over some typical baseline methods.

Abstract

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.

Figures (6)

• Figure 1: Parameter Samples Scaling Law of FISH Mask
• Figure 2: llustrative compared results on SST-2
• Figure 3: Comparison results between FISH Mask method and IRD-based optimization method on BERT-base
• Figure 4: Comparison results between FISH Mask method and IRD-based optimization model on GPT-2
• Figure 5: Contrastive Study