LGAIMay 28, 2025

Adaptive Budget Allocation for Orthogonal-Subspace Adapter Tuning in LLMs Continual Learning

arXiv:2505.22358v21 citationsh-index: 3
Originality Incremental advance
AI Analysis

This addresses the problem of efficient and effective continual learning for LLMs, though it appears incremental by building on existing orthogonal subspace methods.

The paper tackles catastrophic forgetting in continual learning for large language models by proposing OA-Adapter, which unifies dynamic budget adaptation with orthogonal subspace learning in an end-to-end stage, achieving higher average accuracy with 58.5% fewer parameters on a standard benchmark.

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.

Foundations

The foundational work for this paper's niche, ranked by how specifically the neighbourhood builds on it — not by global fame.

Your Notes