Filters








5 Hits in 1.2 sec

secureTF: A Secure TensorFlow Framework [article]

Do Le Quoc, Franz Gregor, Sergei Arnautov, Roland Kunkel, Pramod Bhatotia, Christof Fetzer
2021 arXiv   pre-print
To tackle this challenge, we designed secureTF, a distributed secure machine learning framework based on Tensorflow for the untrusted cloud infrastructure. secureTF is a generic platform to support unmodified  ...  TensorFlow applications, while providing end-to-end security for the input data, ML model, and application code. secureTF is built from ground-up based on the security properties provided by Trusted Execution  ...  To overcome these design challenges, we present secureTF, a secure machine learning framework for the untrusted infrastructure.  ... 
arXiv:2101.08204v1 fatcat:w5zjlifjrfae5az6yia2owbvre

Plinius: Secure and Persistent Machine Learning Model Training [article]

Peterson Yuhala, Pascal Felber, Valerio Schiavoni, Alain Tchana
2021 arXiv   pre-print
We present PLINIUS, a ML framework using Intel SGX enclaves for secure training of ML models and PM for fault tolerance guarantees.  ...  after a system failure.  ...  SecureTF [25] integrates TensorFlow ML library for model training and inference in secure SCONE containers.  ... 
arXiv:2104.02987v2 fatcat:5btty7krfncmdj6adrbl4xqmje

Citadel: Protecting Data Privacy and Model Confidentiality for Collaborative Learning with SGX [article]

Chengliang Zhang, Junzhe Xia, Baichen Yang, Huancheng Puyang, Wei Wang, Ruichuan Chen, Istemi Ekin Akkus, Paarijaat Aditya, Feng Yan
2021 arXiv   pre-print
Cloud deployment with various ML models shows that Citadel scales to a large number of enclaves with less than 1.73X slowdown caused by SGX.  ...  This paper presents Citadel, a scalable collaborative ML system that protects the privacy of both data owner and model owner in untrusted infrastructures with the help of Intel SGX.  ...  We also assume that the participants trust standard ML frameworks like TensorFlow [2] and PyTorch [68] .  ... 
arXiv:2105.01281v2 fatcat:iuc2gbqh4rbpfou2bf7hf7j4em

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

Kha Dinh Duy, Taehyun Noh, Siwon Huh, Hojoon Lee
2021 IEEE Access  
This paper conducts a systematic and comprehensive survey by classifying attack vectors and mitigation in TEE-protected confidential ML computation in the untrusted environment, analyzes the multi-party  ...  ML security requirements, and discusses related engineering challenges.  ...  TensorSCONE [26] introduces TensorFlow to SCONE containers [51] , while secureTF [15] introduces a secure distributed machine learning framework built upon TensorFlow and SGX.  ... 
doi:10.1109/access.2021.3136889 fatcat:scrytvepkjafxblcqg3gjk5vqu

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
This paper conducts a systematic and comprehensive survey by classifying attack vectors and mitigation in TEE-protected confidential ML computation in the untrusted environment, analyzes the multi-party  ...  ML security requirements, and discusses related engineering challenges.  ...  TensorSCONE [26] introduces TensorFlow to SCONE containers [51] , while secureTF [15] introduces a secure distributed machine learning framework built upon TensorFlow and SGX.  ... 
arXiv:2111.03308v2 fatcat:kmklsqvzureilldvr4ui4azrwi