通过双阶段权重保护防御未授权模型合并
Defending Unauthorized Model Merging via Dual-Stage Weight Protection
对做模型合并、多任务权重插值（如 task arithmetic、TIES-merging）的视觉研究者，这篇给出了一个可复用的防御视角：通过 L2 正则化重分布层间任务信息并注入结构化扰动错开任务子空间，使合并后的权重落入破坏性干涉区；反过来也可用于诊断自己合并流程对权重几何的敏感性。
原文依据 we inject structured perturbations to misalign task subspaces, breaking curvature compatibility in the loss landscape
摘要译文
由模型翻译 · 术语统一 · 请对照原文核对关键结论预训练模型与开放仓库的迅速普及使模型合并成为一种便捷却高风险的实践，免费搭车者可在未经授权的情况下将多个微调模型组合成一个新的多能力模型。此类未授权模型合并不仅侵犯知识产权，也损害了模型所有权与可问责性。为解决该问题，我们提出 MergeGuard，一个主动式双阶段权重保护框架，在保持任务保真度的同时破坏合并兼容性。第一阶段通过 L2 正则化优化将任务相关信息在层间重新分配，确保重要梯度被均匀分散。第二阶段注入结构化扰动以错开任务子空间，破坏损失景观中的曲率兼容性。两个阶段共同重塑模型的参数几何，使合并后的模型坍缩为破坏性干涉，而被保护的模型仍保持完整功能。在视觉（ViT-L-14）与语言（Llama2、Gemma2、Mistral）模型上的大量实验表明，MergeGuard 使合并模型的准确率最多下降 90%，而被保护模型自身的性能损失低于 1.5%。
英文原文摘要 1238 字符
The rapid proliferation of pretrained models and open repositories has made model merging a convenient yet risky practice, allowing free-riders to combine fine-tuned models into a new multi-capability model without authorization. Such unauthorized model merging not only violates intellectual property rights but also undermines model ownership and accountability. To address this issue, we present MergeGuard, a proactive dual-stage weight protection framework that disrupts merging compatibility while maintaining task fidelity. In the first stage, we redistribute task-relevant information across layers via L2-regularized optimization, ensuring that important gradients are evenly dispersed. In the second stage, we inject structured perturbations to misalign task subspaces, breaking curvature compatibility in the loss landscape. Together, these stages reshape the model's parameter geometry such that merged models collapse into destructive interference while the protected model remains fully functional. Extensive experiments on both vision (ViT-L-14) and language (Llama2, Gemma2, Mistral) models demonstrate that MergeGuard reduces merged model accuracy by up to 90% with less than 1.5% performance loss on the protected model.