通过分阶段划分领域，提升多领域大模型微调效果。
Synergy over Discrepancy: A Partition-Based Approach to Multi-Domain LLM Fine-Tuning
- 按领域差异与协同效应分阶段微调，减少干扰。
- 在多个语言理解任务上优于当前最优基线。
- 适合需要跨领域适配的大模型应用者。
大型语言模型（LLMs）展现出卓越的泛化能力，但跨异构领域的有效适应仍面临挑战，主要源于域间干扰。为此，我们提出一种基于分区的多阶段微调框架，旨在利用域间协同效应的同时最小化负迁移。该方法通过平衡领域差异、协同效应和模型容量约束，将领域划分为若干子集（阶段）。我们对所提框架进行理论分析，推导出新的泛化误差界，验证了分区策略的合理性。在多种语言理解任务上的大量实证评估表明，该方法持续优于现有最优基线。
原文摘要 · Abstract (English)
Large language models (LLMs) demonstrate impressive generalization abilities, yet adapting them effectively across multiple heterogeneous domains remains challenging due to inter-domain interference. To overcome this challenge, we propose a partition-based multi-stage fine-tuning framework designed to exploit inter-domain synergies while minimizing negative transfer. Our approach strategically partitions domains into subsets (stages) by balancing domain discrepancy, synergy, and model capacity constraints. We theoretically analyze the proposed framework and derive novel generalization bounds that justify our partitioning strategy. Extensive empirical evaluations on various language understanding tasks show that our method consistently outperforms state-of-the-art baselines.
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。