正交子空间适配器微调中的自适应预算分配用于大语言模型持续学习
Adaptive Budget Allocation for Orthogonal-Subspace Adapter Tuning in LLMs Continual Learning
AI总结：
针对大语言模型持续学习中的灾难性遗忘问题，提出OA-Adapter方法，将动态预算分配与正交子空间学习统一于端到端训练，在提升准确率的同时减少58.5%参数。
摘要：
大语言模型（LLMs）在持续学习（CL）场景中常常遭受灾难性遗忘，即在顺序到达的任务上训练时，先前学习任务的性能会严重下降。尽管使用正交子空间的先驱性持续学习方法能够缓解任务干扰，但它们通常采用固定的预算分配，忽略了任务和层之间不同的复杂度。此外，近期针对大语言模型的预算自适应微调方法往往采用多阶段范式，将优化和预算分配解耦。这种解耦导致了潜在的错位，阻碍了这些方法在持续学习场景中的实际应用。为了解决这些局限性，我们提出了OA-Adapter，一种新颖的针对大语言模型持续学习的参数高效方法，它在端到端的训练阶段将动态预算自适应与正交子空间学习统一起来。具体来说，OA-Adapter引入了一种动态瓶颈维度自适应机制，该机制同时分配高效的参数预算并优化任务目标，而不会产生错位。为了在协调动态预算分配的同时有效保留先前获取的知识，正交约束被专门应用于当前任务的参数子空间与历史任务的动态分配参数子空间之间。在持续学习基准上的实验结果表明，OA-Adapter在准确性和参数效率方面均优于最先进的方法。在标准持续学习基准上，OA-Adapter实现了更高的平均准确率，同时使用的参数减少了58.5%，并在包含15个任务的两个更大基准上保持了其优势。
英文摘要：
Large language models (LLMs) often suffer from catastrophic forgetting in continual learning (CL) scenarios, where performance on previously learned tasks degrades severely while training on sequentially arriving tasks. Although pioneering CL approaches using orthogonal subspaces can mitigate task interference, they typically employ fixed budget allocation, neglecting the varying complexity across tasks and layers. Besides, recent budget-adaptive tuning methods for LLMs often adopt multi-stage paradigms that decouple optimization and budget allocation. Such decoupling results in potential misalignment, which hinders those approaches' practical application in CL scenarios. To address these limitations, we propose OA-Adapter, a novel parameter-efficient approach for continual learning in LLMs that unifies dynamic budget adaptation with orthogonal subspace learning in an end-to-end training stage. Specifically, OA-Adapter introduces a dynamic bottleneck dimension adaptation mechanism that simultaneously allocates an efficient parameter budget and optimizes task objectives without misalignment.To effectively preserve previously acquired knowledge while coordinating with the dynamic budget allocation, orthogonal constraints are applied specifically between the parameter subspace of the current task and the dynamically allocated parameter subspaces of historical tasks. Experimental results on continual learning benchmarks demonstrate that OA-Adapter outperforms state-of-the-art methods in both accuracy and parameter efficiency. OA-Adapter achieves higher average accuracy while using 58.5% fewer parameters on the standard CL benchmark, and maintains its advantages on two larger benchmarks comprising 15 tasks.
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。
