多任务模型合并为整合多个微调模型的知识提供了一种高效的解决方案，从而缓解了多任务训练带来的巨大计算和存储需求。作为该领域的一项关键技术，Task Arithmetic (TA) 通过在参数空间中将预训练模型（$θ_{\ ext{pre}}$）从微调任务模型中减去来定义任务向量，然后调整这些任务向量与 $θ_{\ ext{pre}}$ 之间的权重，以平衡任务通用知识和任务特定知识。尽管 TA 性能出色，但任务向量之间可能会出现冲突，特别是当不同任务需要不同的模型适应时。在本文中，我们将此问题正式定义为知识冲突，其特征是在与为另一任务微调的模型合并后，某一任务的性能出现下降。通过深入分析，我们表明这些冲突主要源于任务向量中与 $θ_{\ ext{pre}}$ 处任务特定损失梯度对齐的成分。为了解决这个问题，我们提出了 Task Arithmetic in Trust Region (TATR)，它将信任区域定义为模型参数空间中仅引起任务特定损失微小变化（对应于具有梯度正交方向的任务向量成分）的维度。将参数合并限制在此信任区域内，TATR 可以有效缓解知识冲突。此外，TATR 既可作为一种独立的方法，也可作为兼容多种基于 TA 方法的即插即用模块。在八个不同数据集上的广泛实证评估稳健地证明，TATR 以可观察的幅度提升了多种基于 TA 的模型合并方法的多任务性能。
英文摘要：
Multi-task model merging offers an efficient solution for integrating knowledge from multiple fine-tuned models, mitigating the significant computational and storage demands associated with multi-task training. As a key technique in this field, Task Arithmetic (TA) defines task vectors by subtracting the pre-trained model ($θ_{\text{pre}}$) from the fine-tuned task models in parameter space, then adjusting the weight between these task vectors and $θ_{\text{pre}}$ to balance task-generalized and task-specific knowledge. Despite the promising performance of TA, conflicts can arise among the task vectors, particularly when different tasks require distinct model adaptations. In this paper, we formally define this issue as knowledge conflicts, characterized by the performance degradation of one task after merging with a model fine-tuned for another task. Through in-depth analysis, we show that these conflicts stem primarily from the components of task vectors that align with the gradient of task-specific losses at $θ_{\text{pre}}$. To address this, we propose Task Arithmetic in Trust Region (TATR), which defines the trust region as dimensions in the model parameter space that cause only small changes (corresponding to the task vector components with gradient orthogonal direction) in the task-specific losses. Restricting parameter merging within this trust region, TATR can effectively alleviate knowledge conflicts. Moreover, TATR serves as both an independent approach and a plug-and-play module compatible with a wide range of TA-based methods. Extensive empirical evaluations on eight distinct datasets robustly demonstrate that TATR improves the multi-task performance of several TA-based model merging methods by an observable margin.
Bridging Domains through Subspace-Aware Model Merging
Universidade Estadual de Campinas (UNICAMP)（科林斯州立大学）；Recod.ai Lab.（Recod.ai实验室）；Instituto de Computação（计算研究所）；CISPA Helmholtz Center for Information Security（信息安全赫尔姆霍茨中心）；Intercom
Conflict-Free Replicated Data Types for Neural Network Model Merging: A Two-Layer Architecture Enabling CRDT-Compliant Model Merging Across 26 Strategies
Mediator: Memory-efficient LLM Merging with Less Parameter Conflicts and Uncertainty Based Routing
The Hong Kong University of Science and Technology (Guangzhou)（香港科技大学（广州））；The Hong Kong University of Science and Technology（香港科技大学）；Platform and Content Group, Tencent（腾讯平台与内容事业群）；UCAS（中国科学院大学）；Sun Yat-sen University（中山大学）
Sparsity Curse: Understanding RLVR Model Parameter Space from Model Merging
Zhejiang University（浙江大学）；Simon Fraser University（西蒙菲莎大学）；The Chinese University of Hong Kong（香港中文大学）；Zhejiang Key Lab of Accessible Perception and Intelligent Systems（浙江省可感知智能系统重点实验室）
CALM: Consensus-Aware Localized Merging for Multi-Task Learning
Institute for Artificial Intelligence, Tsinghua University (THUAI)（人工智能研究所，清华大学）；Beijing National Research Center for Information Science and Technology (BNRist)（北京国家信息科学与技术研究中心）；Department of Automation, Tsinghua University（自动化系，清华大学）；East China Normal University (ECNU)（华东师范大学）；The Chinese University of Hong Kong, Shenzhen（香港中文大学（深圳））；School of Computing and Information Systems, The University of Melbourne（墨尔本大学计算机与信息系统学院）
City University of Hong Kong（香港城市大学）；Tencent Jarvis Lab（腾讯 Jarvis 实验室）；University of Technology Sydney（悉尼大学）；University of Queensland（昆士兰大学）；Jilin University（吉林大学）；Westlake University（西湖大学）
Geometry Conflict: Explaining and Controlling Forgetting in LLM Continual Post-Training
The Hong Kong Polytechnic University, PolyU（香港理工大学，PolyU）；The Hong Kong University of Science and Technology, HKUST（香港科学与技术大学，HKUST）；PolyU-Daya Bay Technology and Innovation Research Institute（PolyU-大亚湾技术与创新研究院）
PACT: Preserving Anchored Cores in Task-vectors for Model Merging
Shanghai Jiao Tong University（上海交通大学）；Nanyang Technological University（南洋理工大学）；The Hong Kong University of Science and Technology (Guangzhou)（香港科技大学（广州））
Auto-FlexSwitch: Efficient Dynamic Model Merging via Learnable Task Vector Compression
School of Mathematics, Harbin Institute of Technology, Harbin, P. R. China（哈尔滨工业大学数学学院，哈尔滨，中华人民共和国）；Shanghai Artificial Intelligence Laboratory, Shanghai, P. R. China（上海人工智能实验室，上海，中华人民共和国）
AdaMerging: Adaptive Model Merging for Multi-Task Learning
Northeastern University（东北大学）；University of Oxford（牛津大学）；Nanyang Technological University（南洋理工大学）；University of Maryland（马里兰大学）；JD Explore Academy（京东探索研究院）
