Ownership verification of DNN architectures via hardware cache side channels
Deep Neural Networks (DNN) are gaining higher commercial values in computer vision applications, e.g., image classification, video analytics, etc. This calls for urgent demands of the intellectual property (IP) protection of DNN models. In this paper, we present a novel watermarking scheme to achiev...
| Main Authors: | , , , |
|---|---|
| Other Authors: | |
| Format: | Journal Article |
| Language: | English |
| Published: |
2022
|
| Subjects: | |
| Online Access: | https://hdl.handle.net/10356/159773 |
| _version_ | 1826113727680741376 |
|---|---|
| author | Lou, Xiaoxuan Guo, Shangwei Li, Jiwei Zhang, Tianwei |
| author2 | School of Computer Science and Engineering |
| author_facet | School of Computer Science and Engineering Lou, Xiaoxuan Guo, Shangwei Li, Jiwei Zhang, Tianwei |
| author_sort | Lou, Xiaoxuan |
| collection | NTU |
| description | Deep Neural Networks (DNN) are gaining higher commercial values in computer vision applications, e.g., image classification, video analytics, etc. This calls for urgent demands of the intellectual property (IP) protection of DNN models. In this paper, we present a novel watermarking scheme to achieve the ownership verification of DNN architectures. Existing works all embedded watermarks into the model parameters while treating the architecture as public property. These solutions were proven to be vulnerable by an adversary to detect or remove the watermarks. In contrast, we claim the model architectures as an important IP for model owners, and propose to implant watermarks into the architectures. We design new algorithms based on Neural Architecture Search (NAS) to generate watermarked architectures, which are unique enough to represent the ownership, while maintaining high model usability. Such watermarks can be extracted via side-channel-based model extraction techniques with high fidelity. We conduct comprehensive experiments on watermarked CNN models for image classification tasks and the experimental results show our scheme has negligible impact on the model performance, and exhibits strong robustness against various model transformations and adaptive attacks. |
| first_indexed | 2024-10-01T03:27:58Z |
| format | Journal Article |
| id | ntu-10356/159773 |
| institution | Nanyang Technological University |
| language | English |
| last_indexed | 2024-10-01T03:27:58Z |
| publishDate | 2022 |
| record_format | dspace |
| spelling | ntu-10356/1597732022-07-05T07:20:56Z Ownership verification of DNN architectures via hardware cache side channels Lou, Xiaoxuan Guo, Shangwei Li, Jiwei Zhang, Tianwei School of Computer Science and Engineering Engineering::Computer science and engineering Engineering::Computer science and engineering::Computing methodologies::Artificial intelligence Deep Neural Network Watermarking Cache Side Channel Deep Neural Networks (DNN) are gaining higher commercial values in computer vision applications, e.g., image classification, video analytics, etc. This calls for urgent demands of the intellectual property (IP) protection of DNN models. In this paper, we present a novel watermarking scheme to achieve the ownership verification of DNN architectures. Existing works all embedded watermarks into the model parameters while treating the architecture as public property. These solutions were proven to be vulnerable by an adversary to detect or remove the watermarks. In contrast, we claim the model architectures as an important IP for model owners, and propose to implant watermarks into the architectures. We design new algorithms based on Neural Architecture Search (NAS) to generate watermarked architectures, which are unique enough to represent the ownership, while maintaining high model usability. Such watermarks can be extracted via side-channel-based model extraction techniques with high fidelity. We conduct comprehensive experiments on watermarked CNN models for image classification tasks and the experimental results show our scheme has negligible impact on the model performance, and exhibits strong robustness against various model transformations and adaptive attacks. Ministry of Education (MOE) Nanyang Technological University National Research Foundation (NRF) Submitted/Accepted version This project is in part supported by National Natural Science Foundation of China under Grants U21A20463 and 62102052, Natural Science Foundation of Chongqing, China, under Grant cstc2021jcyj-msxmX0744, Singapore National Research Foundation under its National Cybersecurity R\&D Programme (NCR Award NRF2018NCR-NCR009-0001), Singapore Ministry of Education (MOE) AcRF Tier 2 MOE-T2EP20121-0006, AcRF Tier 1 RS02/19, and NTU Start-up grant. 2022-07-05T07:20:56Z 2022-07-05T07:20:56Z 2022 Journal Article Lou, X., Guo, S., Li, J. & Zhang, T. (2022). Ownership verification of DNN architectures via hardware cache side channels. IEEE Transactions On Circuits and Systems for Video Technology. https://dx.doi.org/10.1109/TCSVT.2022.3184644 1051-8215 https://hdl.handle.net/10356/159773 10.1109/TCSVT.2022.3184644 en NRF2018NCR- NCR009-0001 MOE-T2EP20121-0006 RS02/19 U21A20463 & 62102052 cstc2021jcyj-msxmX0744 IEEE Transactions on Circuits and Systems for Video Technology © 2022 IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting/republishing this material for advertising or promotional purposes, creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works. The published version is available at: https://doi.org/10.1109/TCSVT.2022.3184644. application/pdf |
| spellingShingle | Engineering::Computer science and engineering Engineering::Computer science and engineering::Computing methodologies::Artificial intelligence Deep Neural Network Watermarking Cache Side Channel Lou, Xiaoxuan Guo, Shangwei Li, Jiwei Zhang, Tianwei Ownership verification of DNN architectures via hardware cache side channels |
| title | Ownership verification of DNN architectures via hardware cache side channels |
| title_full | Ownership verification of DNN architectures via hardware cache side channels |
| title_fullStr | Ownership verification of DNN architectures via hardware cache side channels |
| title_full_unstemmed | Ownership verification of DNN architectures via hardware cache side channels |
| title_short | Ownership verification of DNN architectures via hardware cache side channels |
| title_sort | ownership verification of dnn architectures via hardware cache side channels |
| topic | Engineering::Computer science and engineering Engineering::Computer science and engineering::Computing methodologies::Artificial intelligence Deep Neural Network Watermarking Cache Side Channel |
| url | https://hdl.handle.net/10356/159773 |
| work_keys_str_mv | AT louxiaoxuan ownershipverificationofdnnarchitecturesviahardwarecachesidechannels AT guoshangwei ownershipverificationofdnnarchitecturesviahardwarecachesidechannels AT lijiwei ownershipverificationofdnnarchitecturesviahardwarecachesidechannels AT zhangtianwei ownershipverificationofdnnarchitecturesviahardwarecachesidechannels |