Filters








13 Hits in 3.8 sec

Deep Neural Network Fingerprinting by Conferrable Adversarial Examples [article]

Nils Lukas, Yuxuan Zhang, Florian Kerschbaum
2021 arXiv   pre-print
In Machine Learning as a Service, a provider trains a deep neural network and gives many users access.  ...  We propose a new method to generate these conferrable adversarial examples.  ...  INTRODUCTION Deep neural network (DNN) classifiers have become indispensable tools for addressing practically relevant problems, such as autonomous driving (Tian et al., 2018) , natural language processing  ... 
arXiv:1912.00888v4 fatcat:sen7glewxfhv7cjq72zxvyqfqy

Intellectual Property Protection for Deep Learning Models: Taxonomy, Methods, Attacks, and Evaluations [article]

Mingfu Xue, Yushu Zhang, Jian Wang, Weiqiang Liu
2021 arXiv   pre-print
To deal with such security threats, a few deep neural networks (DNN) IP protection methods have been proposed in recent years.  ...  The training and creation of deep learning model is usually costly, thus it can be regarded as an intellectual property (IP) of the model creator.  ...  [40] use conferrable adversarial examples as the fingerprint of DNN model, which is resistant to distillation attacks.  ... 
arXiv:2011.13564v2 fatcat:lbts5q52b5axzlmuitb2u62dfa

A Novel Verifiable Fingerprinting Scheme for Generative Adversarial Networks [article]

Guanlin Li, Guowen Xu, Han Qiu, Shangwei Guo, Run Wang, Jiwei Li, Tianwei Zhang
2021 arXiv   pre-print
This paper presents a novel fingerprinting scheme for the Intellectual Property (IP) protection of Generative Adversarial Networks (GANs).  ...  Prior solutions for classification models adopt adversarial examples as the fingerprints, which can raise stealthiness and robustness problems when they are applied to the GAN models.  ...  CFP-AE Our first fingerprinting method, CFP-AE (Composite Fingerprint based on Adversarial Examples), is inspired by the generative adversarial examples [44] .  ... 
arXiv:2106.11760v2 fatcat:cqrkmazqhrcptmqynpgifj4jjm

Amicable Aid: Turning Adversarial Attack to Benefit Classification [article]

Juyeop Kim, Jun-Ho Choi, Soobeom Jang, Jong-Seok Lee
2021 arXiv   pre-print
While adversarial attacks on deep image classification models pose serious security concerns in practice, this paper suggests a novel paradigm where the concept of adversarial attacks can benefit classification  ...  Furthermore, with a large amount of perturbation, an image can be made unrecognizable by human eyes, while it is correctly recognized by the model.  ...  Deep neural network fingerprinting by conferrable adversarial examples.  ... 
arXiv:2112.04720v1 fatcat:opytok2y2rgehajq2z7jsvurjq

A survey of deep neural network watermarking techniques [article]

Yue Li and Hongxia Wang and Mauro Barni
2021 arXiv   pre-print
Protecting the Intellectual Property Rights (IPR) associated to Deep Neural Networks (DNNs) is a pressing need pushed by the high costs required to train such networks and the importance that DNNs are  ...  In this paper, we overview the most recent advances in DNN watermarking, by paying attention to cast it into the bulk of watermarking theory developed during the last two decades, while at the same time  ...  Introduction Deep Neural Networks (DNNs) are increasingly deployed and commercialised in a wide variety of real-world scenarios due to the unprecedented performance they achieve.  ... 
arXiv:2103.09274v1 fatcat:4volfm2kxfb3zmkgjea62niwpq

ModelDiff: Testing-Based DNN Similarity Comparison for Model Reuse Detection [article]

Yuanchun Li, Ziqi Zhang, Bingyan Liu, Ziyue Yang, Yunxin Liu
2021 pre-print
A study on mobile deep learning apps has shown the feasibility of ModelDiff on real-world models.  ...  In this paper, we propose ModelDiff, a testing-based approach to deep learning model similarity comparison.  ...  INTRODUCTION Deep learning models (i.e. deep neural networks, or DNNs for short) are increasingly deployed into various applications for a wide range of tasks.  ... 
doi:10.1145/3460319.3464816 arXiv:2106.08890v1 fatcat:e2twgte3rrayxann7dturv6r5m

Tracking and Controlling Persons: Identification for Control in Cognitive Behaviours and Cultural Practices

Nicolas J. Bullot
2015
Consider this example from the forensic domain: tracking a killer's organism at a crime scene by means of an analysis of the latent fingerprints and DNA samples left by the killer's organism is a form  ...  For example, this interpretation found its way to a brochure about the history of fingerprinting published by the FBI, where the West case is presented as a "nearly perfect case" that falsifies the claims  ... 
doi:10.7939/r37941833 fatcat:s3vw5adawfcolaszayi6h5ds7m

NSUWorks NSUWorks Citation

Nova
2011 unpublished
BSMP 3170 The Deep Ocean (3 credits) The whole of the ocean environment, down to the very greatest depths at more than 11 kilometers, is populated by living organisms.  ...  Other requirements, which may be held by the affiliated facility include, but are not limited to, fingerprinting, criminal background check, urinalysis for drugs and alcohol, and proof of immunization.  ...  Many argue that video games and music lyrics can cause violence, citing, for example, the Columbine High School mass murders committed by Dylan Klebold and Eric Harris as evidence of what happens when  ... 
fatcat:mwhnd2wtsnd7ti7gvrjwmow3my

NSUWorks NSUWorks Citation

Nova
unpublished
by policies published in a later catalog.  ...  or to requirements approved by the student's academic program director.  ...  BSMP 3170 The Deep Ocean (3 credits) The whole of the ocean environment, down to the very greatest depths at more than 11 kilometers, is populated by living organisms.  ... 
fatcat:w3auuj2sxbfabj6mewlt4ev6lu

NSUWorks NSUWorks Citation

Nova
unpublished
To recklessly or maliciously interfere with or damage computer or network resources or computer data, files, or other information Examples (not a comprehensive list) of policy violations related to the  ...  Student email accounts are deleted upon notification of academic dismissal, suspension, withdrawal, and conferral from various offices of the university.  ...  Nova Southeastern University provides computer systems with access to hardware, software, and networks to enhance academic experience.  ... 
fatcat:bazl2ekuyzdy3mpl73qc2ilvfy

NSUWorks Farquhar College of Arts and Sciences Catalog 2004-2005 (Including information about the undergraduate programs of the Fischler School of Education and Human Services and the

H Wayne
2004 unpublished
networks are investigated.  ...  Minors will be noted on transcripts at the time of degree conferral.  ... 
fatcat:um63kofbfbeybd5ywi4ctuefcq

Nova Southeastern University Farquhar College of Arts and Sciences Catalog

Nova, H Wayne
2004 unpublished
networks are investigated.  ...  Minors will be noted on transcripts at the time of degree conferral.  ... 
fatcat:gukhliukrbgqveq47424sbk33e

NSUWorks NSUWorks Citation

Nova
2008 unpublished
Other requirements, which may be held by the affiliated facility include, but are not limited to, fingerprinting, criminal background check, urinalysis for drugs and alcohol, and proof of immunization.  ...  A minor will be posted to the student's transcript at the time of conferral of the bachelor's degree. Minors will not be posted following conferral of the degree.  ...  HONR 2000R Honors Seminar: The Misbehaving Brain (3 credits) Students will develop an understanding of the neural correlates and behavioral impairments seen in different neuropsychiatric and neurological  ... 
fatcat:tmmtudhgozapxe7jcwyd7lhj2e