Cover Image

DeepIPR: Deep neural network ownership verification with passports

With substantial amount of time, resources and human (team) efforts invested to explore and develop successful deep neural networks (DNN), there emerges an urgent need to protect these inventions from being illegally copied, redistributed, or abused without respecting the intellectual properties of...

Full description

Saved in:
Bibliographic Details
Main Authors: Fan, Lixin, Ng, Kam Woh, Chan, Chee Seng, Yang, Qiang
Format: Article
Published: IEEE Computer Soc 2022
Subjects:
QA75 Electronic computers. Computer science
Online Access:http://eprints.um.edu.my/41296/
Tags: Add Tag
No Tags, Be the first to tag this record!
id my.um.eprints.41296
record_format eprints
spelling my.um.eprints.412962023-09-18T04:11:27Z http://eprints.um.edu.my/41296/ DeepIPR: Deep neural network ownership verification with passports Fan, Lixin Ng, Kam Woh Chan, Chee Seng Yang, Qiang QA75 Electronic computers. Computer science With substantial amount of time, resources and human (team) efforts invested to explore and develop successful deep neural networks (DNN), there emerges an urgent need to protect these inventions from being illegally copied, redistributed, or abused without respecting the intellectual properties of legitimate owners. Following recent progresses along this line, we investigate a number of watermark-based DNN ownership verification methods in the face of ambiguity attacks, which aim to cast doubts on the ownership verification by forging counterfeit watermarks. It is shown that ambiguity attacks pose serious threats to existing DNN watermarking methods. As remedies to the above-mentioned loophole, this paper proposes novel passport-based DNN ownership verification schemes which are both robust to network modifications and resilient to ambiguity attacks. The gist of embedding digital passports is to design and train DNN models in a way such that, the DNN inference performance of an original task will be significantly deteriorated due to forged passports. In other words, genuine passports are not only verified by looking for the predefined signatures, but also reasserted by the unyielding DNN model inference performances. Extensive experimental results justify the effectiveness of the proposed passport-based DNN ownership verification schemes. Code is available at https://github.com/kamwoh/DeepIPR. IEEE Computer Soc 2022-10 Article PeerReviewed Fan, Lixin and Ng, Kam Woh and Chan, Chee Seng and Yang, Qiang (2022) DeepIPR: Deep neural network ownership verification with passports. IEEE Transactions on Pattern Analysis and Machine Intelligence, 44 (10). pp. 6122-6139. ISSN 0162-8828, DOI https://doi.org/10.1109/TPAMI.2021.3088846 <https://doi.org/10.1109/TPAMI.2021.3088846>. 10.1109/TPAMI.2021.3088846
institution Universiti Malaya
building UM Library
collection Institutional Repository
continent Asia
country Malaysia
content_provider Universiti Malaya
content_source UM Research Repository
url_provider http://eprints.um.edu.my/
topic QA75 Electronic computers. Computer science
spellingShingle QA75 Electronic computers. Computer science
Fan, Lixin
Ng, Kam Woh
Chan, Chee Seng
Yang, Qiang
DeepIPR: Deep neural network ownership verification with passports
description With substantial amount of time, resources and human (team) efforts invested to explore and develop successful deep neural networks (DNN), there emerges an urgent need to protect these inventions from being illegally copied, redistributed, or abused without respecting the intellectual properties of legitimate owners. Following recent progresses along this line, we investigate a number of watermark-based DNN ownership verification methods in the face of ambiguity attacks, which aim to cast doubts on the ownership verification by forging counterfeit watermarks. It is shown that ambiguity attacks pose serious threats to existing DNN watermarking methods. As remedies to the above-mentioned loophole, this paper proposes novel passport-based DNN ownership verification schemes which are both robust to network modifications and resilient to ambiguity attacks. The gist of embedding digital passports is to design and train DNN models in a way such that, the DNN inference performance of an original task will be significantly deteriorated due to forged passports. In other words, genuine passports are not only verified by looking for the predefined signatures, but also reasserted by the unyielding DNN model inference performances. Extensive experimental results justify the effectiveness of the proposed passport-based DNN ownership verification schemes. Code is available at https://github.com/kamwoh/DeepIPR.
format Article
author Fan, Lixin
Ng, Kam Woh
Chan, Chee Seng
Yang, Qiang
author_facet Fan, Lixin
Ng, Kam Woh
Chan, Chee Seng
Yang, Qiang
author_sort Fan, Lixin
title DeepIPR: Deep neural network ownership verification with passports
title_short DeepIPR: Deep neural network ownership verification with passports
title_full DeepIPR: Deep neural network ownership verification with passports
title_fullStr DeepIPR: Deep neural network ownership verification with passports
title_full_unstemmed DeepIPR: Deep neural network ownership verification with passports
title_sort deepipr: deep neural network ownership verification with passports
publisher IEEE Computer Soc
publishDate 2022
url http://eprints.um.edu.my/41296/
_version_ 1778161652601978880
score 13.211869

Similar Items