Octopus：面向多模态大语言模型持续学习的无历史梯度正交化
Octopus: History-Free Gradient Orthogonalization for Continual Learning in Multimodal Large Language Models
为视觉-语言模型的多任务顺序微调提供了一种无需存储历史数据的梯度正交化约束，可直接用于缓解 MLLM 在连续下游视觉任务（如 VQA、captioning）微调时的灾难性遗忘。
原文依据 enforces gradient-level orthogonality without historical task data
摘要译文
由模型翻译 · 术语统一 · 请对照原文核对关键结论多模态大语言模型（MLLM）中的持续学习旨在顺序获取知识的同时缓解灾难性遗忘，然而现有方法存在固有局限：基于架构的方法会带来额外计算开销，且往往难以泛化到新任务；基于回放的方法依赖存储历史数据，引发隐私与存储方面的顾虑；而仅靠传统正则化策略不足以完全防止参数干扰。我们提出 Octopus，一个基于无历史梯度正交化（HiFGO）的两阶段持续学习框架，在无需历史任务数据的情况下实现梯度级正交性。该两阶段微调策略将任务适配与正则化解耦，在可塑性与稳定性之间取得有原则的平衡。在 UCIT 上的实验表明，Octopus 达到最优性能，在 Avg 和 Last 指标上分别超越此前最优方法 2.14% 和 6.82%。
英文原文摘要 997 字符
Continual learning in multimodal large language models (MLLMs) aims to sequentially acquire knowledge while mitigating catastrophic forgetting, yet existing methods face inherent limitations: architecture-based approaches incur additional computational overhead and often generalize poorly to new tasks, rehearsal-based methods rely on storing historical data, raising privacy and storage concerns, and conventional regularization-based strategies alone are insufficient to fully prevent parameter interference. We propose Octopus, a two-stage continual learning framework based on History-Free Gradient Orthogonalization (HiFGO), which enforces gradient-level orthogonality without historical task data. The two-stage finetuning strategy decouples task adaptation from regularization, achieving a principled balance between plasticity and stability. Experiments on UCIT show that Octopus establishes state-of-the-art performance, surpassing prior SOTA by 2.14% and 6.82% in terms of Avg and Last.