在开放测试流中学习并适应未知
Toward Robust Open-set Adaptation: Synapse Consolidation Inspired by Rac1/MAPK Pathways
AI总结：
本文提出SyCo方法，通过Rac1和MAPK路径实现参数高效适应，解决LLM在动态任务中的适应问题，实验表明其在18个NLP数据集上表现优异。
摘要：
大型语言模型（LLMs）通过可重用的表示和灵活的推理跨任务泛化，但在实际部署中面对演变任务和持续分布偏移时仍易碎。一种常见方法是测试时适应（TTA），现有方法通过手设计的无监督目标更新模型，大多忽视保留共享源知识和适应信号的可靠性。受果蝇记忆更新分子信号级联的启发，我们提出Synapse Consolidation（SyCo），一种参数高效的LLM适应方法，通过Rac1和MAPK路径在结构化TTA目标指导下更新低秩适配器。Rac1将可塑性限制在尾梯度子空间，该空间对源知识较少关键，从而实现快速专业化同时保留源表示。MAPK使用分层控制器抑制噪声更新并在非平稳流中巩固有用适应。为建模具有多个源和持续新兴任务的现实部署，我们引入多源开放集适应（MOA）设置，其中模型在多个标记源任务上训练，然后在开放、非平稳的未标记测试流中适应，该流混合已见和未见任务，部分在标签和意图空间上有重叠。在18个NLP数据集和MOA设置上，SyCo一致优于强基线，实现78.31%的未见任务适应和85.37%的未见数据偏移。
英文摘要：
Large Language Models (LLMs) generalize across tasks through reusable representations and flexible reasoning, yet remain brittle in real deployment when faced with evolving tasks and continual distribution shift. While test-time adaptation addresses this by updating models with unsupervised objectives on test data, prevailing methods are fundamentally limited by their neglect of source knowledge preservation and adaptation signal reliability. Inspired by how Drosophila orchestrates memory update by balancing retroactive and proactive interference via Rac1 and MAPK pathways, we design Synapse Consolidation (SyCo) with two core components: a Rac1-inspired plasticity confiner and a MAPK-inspired update controller. The former dynamically confines plasticity to a tail-gradient subspace that is less critical for source knowledge, enabling rapid specialization while preserving source representations. The latter uses a tiered controller to suppress noisy updates and consolidate useful adaptations under non-stationary streams. To further model real deployments with multiple sources and continually emerging tasks, we introduce Multi-source Open-set Adaptation (MOA) setting, where a model is trained on multiple labeled source tasks and then adapts on open, non-stationary unlabeled test streams mixing seen and unseen tasks with partial overlap in label and intent space. Across 18 NLP datasets under the MOA setting, SyCo consistently outperforms strong baselines, achieving 78.31\% on unseen-task adaptation and 85.37\% versus unseen-data shifts, setting a new state-of-the-art.
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。
