7 Hits in 2.7 sec

Faster CryptoNets: Leveraging Sparsity for Real-World Encrypted Inference [article]

Edward Chou, Josh Beal, Daniel Levy, Serena Yeung, Albert Haque, Li Fei-Fei
2018 arXiv   pre-print
We present Faster CryptoNets, a method for efficient encrypted inference using neural networks.  ...  We also show how privacy-safe training techniques can be used to reduce the overhead of encrypted inference for real-world datasets by leveraging transfer learning and differential privacy.  ...  Contributions We propose Faster CryptoNets -a method for encrypted inference on the order of seconds.  ... 
arXiv:1811.09953v1 fatcat:67vqy4vwqvahvd6kqcxuzp23ki

SoK: Privacy-preserving Deep Learning with Homomorphic Encryption [article]

Robert Podschwadt, Daniel Takabi, Peizhao Hu
2022 arXiv   pre-print
With homomorphic encryption (HE) computation can be performed on encrypted data without revealing its content.  ...  In this systematization of knowledge, we take an in-depth look at approaches that combine neural networks with HE for privacy preservation.  ...  for real-world encrypted inference,” Medical Association.  ... 
arXiv:2112.12855v2 fatcat:35nloctuj5ahlgnlze3u2mcgz4

CryptoNite: Revealing the Pitfalls of End-to-End Private Inference at Scale [article]

Karthik Garimella, Nandan Kumar Jha, Zahra Ghodsi, Siddharth Garg, Brandon Reagen
2021 arXiv   pre-print
The privacy concerns of providing deep learning inference as a service have underscored the need for private inference (PI) protocols that protect users' data and the service provider's model using cryptographic  ...  Recently proposed PI protocols have achieved significant reductions in PI latency by moving the computationally heavy homomorphic encryption (HE) parts to an offline/pre-compute phase.  ...  For instance, Faster-CryptoNets and HEMET have only low-degree polynomial activations and the former reduces the number of HE operations by introducing the sparsity in the network (using pruning and quantization  ... 
arXiv:2111.02583v1 fatcat:w5cft4qgvrcuhhcfo4nrk57xye

XONN: XNOR-based Oblivious Deep Neural Network Inference [article]

M. Sadegh Riazi and Mohammad Samragh and Hao Chen and Kim Laine and Kristin Lauter and Farinaz Koushanfar
2019 arXiv   pre-print
Advancements in deep learning enable cloud servers to provide inference-as-a-service for clients.  ...  State-of-the-art frameworks require one round of interaction between the client and the server for each layer of the neural network, whereas, XONN requires a constant round of interactions for any number  ...  Acknowledgements We would like to thank the anonymous reviewers for their insightful comments.  ... 
arXiv:1902.07342v2 fatcat:gf5gyhmltbdq5gllo5lckw765y

Scalable privacy-preserving cancer type prediction with homomorphic encryption [article]

Esha Sarkar, Eduardo Chielle, Gamze Gursoy, Leo Chen, Mark Gerstein, Michail Maniatakos
2022 arXiv   pre-print
In this work, we explore the challenges of privacy preserving cancer detection using a real-world dataset consisting of more than 2 million genetic information for several cancer types.  ...  We develop a solution for privacy preserving cancer inference which first leverages the domain knowledge on somatic mutations to efficiently encode genetic mutations and then uses statistical tests for  ...  In summary, to enable practical real-world private inference, we need the ability to compute on high-dimensional data in the encrypted domain with low latency and high throughput.  ... 
arXiv:2204.05496v1 fatcat:rby2isau4zglvkaclus7zewzge

Identifying and Exploiting Structures for Reliable Deep Learning [article]

Amartya Sanyal
2021 arXiv   pre-print
However, as recent works point out, these systems suffer from several issues that make them unreliable for use in the real world, including vulnerability to adversarial attacks (Szegedy et al. [248]),  ...  The extraordinary performance of these systems often gives the impression that they can be used to revolutionise our lives for the better.  ...  for deploying learning algorithms in the real world.  ... 
arXiv:2108.07083v1 fatcat:lducrn5tlfeqvpxevz6gukfvse

Compiler and runtime systems for homomorphic encryption and graph processing on distributed and heterogeneous architectures

Roshan Dathathri, Austin, The University Of Texas At, Keshav Pingali
Fully Homomorphic Encryption (FHE) refers to a set of encryption schemes that allow computations on encrypted data without requiring a secret key.  ...  I present CHET, a domain-specific optimizing compiler, that is designed to make the task of programming DNN inference applications using FHE easier.  ...  Gluon's BASP-style execution is on average ∼ 1.5× faster than its BSP-style execution for graph applications on real-world large-diameter graphs at scale.  ... 
doi:10.26153/tsw/10165 fatcat:cyejg4vtojgsfiwbrbqima6idm