Filters








9 Hits in 3.1 sec

MLCapsule: Guarded Offline Deployment of Machine Learning as a Service [article]

Lucjan Hanzlik, Yang Zhang, Kathrin Grosse, Ahmed Salem, Max Augustin, Michael Backes, Mario Fritz
2019 arXiv   pre-print
In this paper, we propose MLCapsule, a guarded offline deployment of machine learning as a service.  ...  With the widespread use of machine learning (ML) techniques, ML as a service has become increasingly popular.  ...  Our Contributions We propose MLCapsule, a guarded offline deployment of machine learning as a service.  ... 
arXiv:1808.00590v2 fatcat:vm3vc46zjvcerghohniz4gikpq

Offline Model Guard: Secure and Private ML on Mobile Devices

Sebastian P. Bayerl, Tommaso Frassetto, Patrick Jauernig, Korbinian Riedhammer, Ahmad-Reza Sadeghi, Thomas Schneider, Emmanuel Stapf, Christian Weinert
2020 2020 Design, Automation & Test in Europe Conference & Exhibition (DATE)  
Specifically, we build Offline Model Guard (OMG) to enable privacy-preserving machine learning on the predominant mobile computing platform ARM - even in offline scenarios.  ...  Performing machine learning tasks in mobile applications yields a challenging conflict of interest: highly sensitive client information (e.g., speech data) should remain private while also the intellectual  ...  In this work, we build OFFLINE MODEL GUARD (OMG), a generic architecture that efficiently protects machine learning tasks on mobile devices like smartphones and tablets, and demonstrate its practicality  ... 
doi:10.23919/date48585.2020.9116560 dblp:conf/date/BayerlFJRS0SW20 fatcat:wvjhalug6zfgtjdt6ujy2sf7om

Confidential Inference via Ternary Model Partitioning [article]

Zhongshu Gu, Heqing Huang, Jialong Zhang, Dong Su, Hani Jamjoom, Ankita Lamba, Dimitrios Pendarakis, Ian Molloy
2020 arXiv   pre-print
Today's cloud vendors are competing to provide various offerings to simplify and accelerate AI service deployment.  ...  Our research prototype consists of two co-operative components: (1) Model Assessment Framework, a local model evaluation and partitioning tool that assists cloud users in deployment preparation; (2) Infenclave  ...  MLCapsule [15] is another interesting offline model deployment approach that executes model locally on the client's machine and protects the models' secrecy with SGX enclaves.  ... 
arXiv:1807.00969v3 fatcat:y5fxdsexh5dwdklaqg62gj5hxy

Confidential Machine Learning Computation in Untrusted Environments: A Systems Security Perspective [article]

Kha Dinh Duy, Taehyun Noh, Siwon Huh, Hojoon Lee
2021 arXiv   pre-print
As machine learning (ML) technologies and applications are rapidly changing many domains of computing, security issues associated with ML are also emerging.  ...  In the domain of systems security, many endeavors have been made to ensure ML model and data confidentiality.  ...  contributors and model owners) We use the term multi-party machine learning as a service (MPMLaaS) to refer to multi-party computation scenarios in which the service provider provides a service based  ... 
arXiv:2111.03308v2 fatcat:kmklsqvzureilldvr4ui4azrwi

Confidential Machine Learning Computation in Untrusted Environments: A Systems Security Perspective

Kha Dinh Duy, Taehyun Noh, Siwon Huh, Hojoon Lee
2021 IEEE Access  
As machine learning (ML) technologies and applications are rapidly changing many domains of computing, security issues associated with ML are also emerging.  ...  In the domain of systems security, many endeavors have been made to ensure ML model and data confidentiality.  ...  contributors and model owners) We use the term multi-party machine learning as a service (MPMLaaS) to refer to multi-party computation scenarios in which the service provider provides a service based  ... 
doi:10.1109/access.2021.3136889 fatcat:scrytvepkjafxblcqg3gjk5vqu

Modelling of Received Signals in Molecular Communication Systems based machine learning: Comparison of azure machine learning and Python tools [article]

Soha Mohamed, Mahmoud S. Fayed
2021 arXiv   pre-print
Machine learning (ML) is one of the intelligent methodologies that has shown promising results in the domain.  ...  This paper applies Azure Machine Learning (Azure ML) for flexible pavement maintenance regressions problems and solutions.  ...  Hanzlik, L., Zhang, Y., Grosse, K., Salem, A., Augustin, M., Backes, M., & Fritz, M. (2021). Mlcapsule: Guarded offline deployment of machine learning as a service.  ... 
arXiv:2112.10214v1 fatcat:fxnqhqhnjfhtfnfrv2pby5p6lu

Survey of Attacks and Defenses on Edge-Deployed Neural Networks [article]

Mihailo Isakov, Vijay Gadepally, Karen M. Gettings, Michel A. Kinsy
2019 arXiv   pre-print
In this work, we cover the landscape of attacks on, and defenses, of neural networks deployed in edge devices and provide a taxonomy of attacks and defenses targeting edge DNNs.  ...  While datacenter networks can be protected using conventional cybersecurity measures, edge neural networks bring a host of new security challenges.  ...  In MLCapsule [84] , authors develop a machine learning as a service (MLaaS) platform above Trusted Execution Environments (TEE) such as Intel SGX, and formally prove it's security.  ... 
arXiv:1911.11932v1 fatcat:zihiqvq2tbd3zpuyvwqrrf5itq

Privado: Practical and Secure DNN Inference with Enclaves [article]

Karan Grover, Shruti Tople, Shweta Shinde, Ranjita Bhagwan and Ramachandran Ramjee
2019 arXiv   pre-print
In this paper, we ask a timely question: "Can third-party cloud services use Intel SGX enclaves to provide practical, yet secure DNN Inference-as-a-service?"  ...  Cloud providers are extending support for trusted hardware primitives such as Intel SGX. Simultaneously, the field of deep learning is seeing enormous innovation as well as an increase in adoption.  ...  Similarly, MLCapsule, a system for secure but offline deployment of ML as a service on the client-side is susceptible to leakage of sensitive inputs via access patterns [8] .  ... 
arXiv:1810.00602v2 fatcat:aomf6hdikjf5dhgpsjesydnq4e

Why is Machine Learning Security so hard? [article]

Kathrin Grosse, Universität Des Saarlandes
2021
available data and computing power has fueled a wide application of machine learning (ML).  ...  A different kind of complexity is added with the large libraries nowadays in use to implement machine learning.  ...  I am grateful for his support, skills and feedback throughout the time of my PhD studies.  ... 
doi:10.22028/d291-34355 fatcat:tpcmb6zftrflzefiuqyasy5n5e