Model Barrier: A Compact Un-Transferable Isolation Domain for Model Intellectual Property Protection
Lianyu Wang, Meng Wang, Daoqiang Zhang, Huazhu Fu
摘要
As scientific and technological advancements result from human intellectual labor and computational costs, protecting model intellectual property (IP) has become increasingly important to encourage model creators and owners. Model IP protection involves preventing the use of welltrained models on unauthorized domains. To address this issue, we propose a novel approach called Compact Un-Transferable Isolation Domain (CUTI-domain), which acts as a barrier to block illegal transfers from authorized to unauthorized domains. Specifically, CUTI-domain blocks cross-domain transfers by highlighting the private style features of the authorized domain, leading to recognition failure on unauthorized domains with irrelevant private style features. Moreover, we provide two solutions for using CUTI-domain depending on whether the unauthorized domain is known or not: target-specified CUTI-domain and target-free CUTI-domain. Our comprehensive experimental results on four digit datasets, CIFAR10 & STL10, and VisDA-2017 dataset demonstrate that CUTI-domain can be easily implemented as a plug-and-play module with different backbones, providing an efficient solution for model IP protection.
问问这篇 Paper
智能体会读完全文。
Lune 把这篇 Paper 索引到了每一个公式,引用它的顶会 Paper 也一样。你提问,回答直接引用原文。
引用它的顶会 Paper12
- FedTGP: Trainable Global Prototypes with Adaptive-Margin-Enhanced Contrastive Learning for Data and Model Heterogeneity in Federated LearningJianqing Zhang, Yang Liu, Yang Hua, Jian CaoAAAI 2024 · 被引用 142 次
- An Upload-Efficient Scheme for Transferring Knowledge From a Server-Side Pre-trained Generator to Clients in Heterogeneous Federated LearningJianqing Zhang, Yang Liu, Yang Hua, Jian CaoCVPR 2024 · 被引用 24 次
- Sophon: Non-Fine-Tunable Learning to Restrain Task Transferability For Pre-trained ModelsJiangyi Deng, Shengyuan Pang, Yanjiao Chen, Liangming Xia 等S&P 2024 · 被引用 18 次
- FedSA: A Unified Representation Learning via Semantic Anchors for Prototype-based Federated LearningYanbing Zhou, Xiangmou Qu, Chenlong You, Jiyang Zhou 等AAAI 2025 · 被引用 9 次
- Domain Specified Optimization for Deployment AuthorizationHaotian Wang, Haoang Chi, Wenjing Yang, Zhipeng Lin 等ICCV 2023 · 被引用 8 次
它引用的顶会 Paper9
- Swin Transformer: Hierarchical Vision Transformer using Shifted WindowsZe Liu, Yutong Lin, Yue Cao, Han Hu 等ICCV 2021 · 被引用 31,683 次
- Turning Your Weakness Into a Strength: Watermarking Deep Neural Networks by BackdooringYossi Adi, Carsten Baum, Moustapha Cissé, Benny Pinkas 等USENIX Security 2018 · 被引用 832 次
- Machine Learning Models that Remember Too MuchCongzheng Song, Thomas Ristenpart, Vitaly ShmatikovCCS 2017 · 被引用 582 次
- IDM: An Intermediate Domain Module for Domain Adaptive Person Re-IDYongxing Dai, Jun Liu, Yifan Sun, Zekun Tong 等ICCV 2021 · 被引用 145 次
- Hardware-Assisted Intellectual Property Protection of Deep Learning ModelsAbhishek Chakraborty, Ankit Mondal, Ankur SrivastavaDAC 2020 · 被引用 75 次
相关 Paper
- Non-Transferable Learning: A New Approach for Model Ownership Verification and Applicability AuthorizationLixu Wang, Shichao Xu, Ruiqi Xu, Xiao Wang 等ICLR 2022 · 被引用 65 次
- Vision-Language Model IP Protection via Prompt-based LearningLianyu Wang, Meng Wang, Huazhu Fu, Daoqiang ZhangCVPR 2025
- Jailbreaking the Non-Transferable Barrier via Test-Time Data DisguisingYongli Xiang, Ziming Hong, Lina Yao, Dadong Wang 等CVPR 2025
- Authorize-on-Demand: Dynamic Authorization with Legality-Aware Intellectual Property Protection for VLMsLianyu Wang, Meng Wang, Huazhu Fu, Daoqiang ZhangCVPR 2026
- Your Transferability Barrier is Fragile: Free-Lunch for Transferring the Non-Transferable LearningZiming Hong, Li Shen, Tongliang LiuCVPR 2024
