让大模型多任务学习更省数据，仅用少量样本就能高效适配。
MeTA-LoRA: Data-Efficient Multi-Task Fine-Tuning for Large Language Models
- 分两阶段训练：先快速学各任务专用适配器，再共享更新知识。
- 多任务和多语言场景下，用更少数据达到甚至超越全量微调效果。
- 适合资源有限但需多任务部署的场景，如小样本应用开发。
低秩适应（LoRA）是当前广泛使用的参数高效微调方法，适用于单任务场景，但在复杂多任务学习中难以有效利用任务间知识，通常需要大量任务特定数据才能达到最佳性能。为此，我们提出MeTA-LoRA，一种两阶段优化框架，显著提升多任务适配的数据效率。第一阶段仅使用每个数据集的少量样本训练任务专属的LoRA适配器，实现快速适应而无需大规模标注；第二阶段通过聚合多个任务的梯度，更新共享的LoRA适配器，促进任务间知识迁移，进一步减少数据需求。在多任务学习与多语言学习场景中，该方法在显著降低任务特定数据使用量的同时，性能达到或超过传统全数据LoRA微调水平。
原文摘要 · Abstract (English)
Low-Rank Adaptation (LoRA) has emerged as one of the most widely used parameter-efficient fine-tuning (PEFT) methods for adapting large language models (LLMs) to downstream tasks. While highly effective in single-task settings, it struggles to efficiently leverage inter-task knowledge in complex multi-task learning scenarios, often requiring substantial task-specific data to achieve optimal performance. To address this limitation, we introduce MeTA-LoRA, a two-stage optimization framework that significantly improves data efficiency in multi-task adaptation. In the first stage, task-specific LoRA adapters are learned using only a few samples from each involved dataset, enabling rapid adaptation without large-scale supervision. In the second stage, the shared LoRA adapter is updated by aggregating gradients from multiple tasks to promote knowledge transfer across tasks, further reducing data usage by leveraging common patterns. In both multi-task learning and multilingual learning scenarios, our method matches or surpasses the performance of traditional full-data LoRA fine-tuning approaches, while using significantly less task-specific data.
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。