适应大语言模型到新领域会导致遗忘，因为标准方法（如全微调、LoRA）向权重空间注入新方向。我们证明遗忘由一个代数性质决定：更新是否保持预训练权重矩阵的列空间（命题1）。我们提出GAIN，最简单的乘法替代方法（W_new = S * W），通过构造满足这一性质，并可融入现有权重以实现零推理成本。在五种模型（774M至70B）依次适应八个领域的情况下，GAIN将早期领域困惑度提升7-13%，而LoRA下降18-36%。GAIN在不存储先前数据的情况下匹配增强的LoRA，并在遗忘-适应帕累托前沿上优于EWC。虽然LoRA只能通过牺牲领域适应性来减少遗忘，GAIN在无领域边界和无正则化的情况下实现两者。该原理可推广：(IA)^3，一种独立的乘法方法，也提升了早期领域。
英文摘要：
Adapting LLMs to new domains causes forgetting because standard methods (e.g., full fine-tuning, LoRA) inject new directions into the weight space. We show that forgetting is governed by one algebraic property: whether the update preserves the column span of the pretrained weight matrix (Proposition 1). We propose GAIN, the simplest multiplicative alternative (W_new = S * W), which satisfies this by construction and can be absorbed into existing weights for zero inference cost.
Across five models (774M to 70B) adapted sequentially over eight domains, GAIN improves earlier-domain perplexity by 7-13%, while LoRA degrades it by 18-36%. GAIN matches replay-augmented LoRA without storing prior data and dominates EWC on the forgetting-adaptation Pareto front. While LoRA can only reduce forgetting by sacrificing in-domain adaptation, GAIN achieves both with no domain boundaries and no regularization. The principle generalises: (IA)^3, an independent multiplicative method, also improves earlier domains.
Threshold Modulation for Online Test-Time Adaptation of Spiking Neural Networks
Department of CSE, Southern University of Science and Technology（南方科技大学计算机科学与工程系）；College of Mathematics and Statistics, Chongqing University（重庆大学数学与统计学院）；ACS Laboratory, Huawei Technologies Co., Ltd.（华为技术有限公司ACS实验室）
Contextual Attention Modulation: Towards Efficient Multi-Task Adaptation in Large Language Models
SCSE, Beihang University（北京航空航天大学系统工程学院）；ERC of ACAT, MOE（教育部人工智能教育联合体）；City University of Hong Kong（香港城市大学）；Huawei Technologies Ltd.（华为技术有限公司）；Key Lab of DIM, MIIT（信息通信技术重点实验室）；SEM, Beihang University（北京航空航天大学电子机械学院）；Zhejiang University of Technology（浙江工业大学）
EventOD: Event-Aware OD Flow Generation via LLM-Guided Semantic Modulation
Department of Electronic Engineering, Tsinghua University（清华大学电子工程系）；Zhongguancun Academy（中关村科学城研究院）；Singapore-MIT Alliance for Research and Technology（新加坡-麻省理工学院科研与技术联盟）；Xiuzhong College, Tsinghua University（清华大学致理书院）；Department of Automation, Central South University（中南大学自动化学院）
Breaking the Blocks: Continuous Low-Rank Decomposed Scaling for Unified LLM Quantization and Adaptation
Institute for Artificial Intelligence, Peking University, Beijing, China（北京大学人工智能研究院）；Yuanpei College, Peking University, Beijing, China（北京大学元培学院）；State Key Laboratory of General Artificial Intelligence, BIGAI, Beijing, China（通用人工智能国家重点实验室）；Harbin Institute of Technology, Shenzhen, China（哈尔滨工业大学（深圳））
Parameter-Efficient Vision-Language Adaptation with Continuous Metadata Conditioning for Animal Re-Identification
Department of Computer Science, University of Verona（维罗纳大学计算机科学系）；Institute of Marine Research（海洋研究所）；University of Agder, Centre for Coastal Research（阿格德大学海岸研究中心）
AI Knows When It's Being Watched: Functional Strategic Action and Contextual Register Modulation in Large Language Models
Center for Applied Communication Research (CICA)（应用沟通研究中心）；Human & NonHuman Communication Laboratory（人类与非人类沟通实验室）；Faculty of Communication（传播学院）；Universidad Anáhuac México（墨西哥安纳胡阿克大学）
Adaptive whitening with fast gain modulation and slow synaptic plasticity
Center for Computational Neuroscience, Flatiron Institute（Flatiron Institute计算神经科学中心）；Center for Neural Science, New York University（纽约大学神经科学中心）；Neuroscience Institute, NYU Langone Medical School（纽约大学朗格尼医学院神经科学研究所）
