优化指令学习中的交互与依赖关系，提升大模型效果
Beyond IID: Optimizing Instruction Learning from the Perspective of Instruction Interaction and Dependency
- 分析不同指令类别间的交互与依赖模式
- 用线性规划优化指令集，提升模型性能
- 基于依赖关系设计课程学习策略，适合大模型训练研究者
随着各类指令数据集的出现，如何有效选择和整合指令以微调大语言模型（LLMs）成为关键挑战。以往研究多聚焦于单个高质量指令的选择，忽略了不同类别指令之间的联合交互与依赖关系，导致选择策略不优。此外，这些交互模式的本质仍缺乏探索，更未被用于优化指令集。为此，本文系统研究了不同类别指令间的交互与依赖模式，提出基于线性规划的方法优化指令集，并采用依赖分类引导的课程学习优化SFT学习方案。在多个LLM和广泛使用的基准测试上，实验结果表明该方法优于强基线。
原文摘要 · Abstract (English)
With the availability of various instruction datasets, a pivotal challenge is how to effectively select and integrate these instructions to fine-tune large language models (LLMs). Previous research mainly focuses on selecting individual high-quality instructions. However, these works overlooked the joint interactions and dependencies between different categories of instructions, leading to suboptimal selection strategies. Moreover, the nature of these interaction patterns remains largely unexplored, let alone optimize the instruction set with regard to them. To fill these gaps, in this paper, we: (1) systemically investigate interaction and dependency patterns between different categories of instructions, (2) manage to optimize the instruction set concerning the interaction patterns using a linear programming-based method, and optimize the learning schema of SFT using an instruction dependency taxonomy guided curriculum learning. Experimental results across different LLMs demonstrate improved performance over strong baselines on widely adopted benchmarks.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。