DArL: Dynamic Parameter Adjustment for LWE-based Secure Inference

Bian, Song; Hiromoto, Masayuki; Satō, Takashi

doi:10.23919/date.2019.8715110

Cited by 6 publications

(13 citation statements)

References 21 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Subsequent work by the authors [80] presented an approximate bootstrapping operation for homomorphic decryption. Also, [81] utilized the approximate computing techniques proposed in [75] to improve the efficiency of homomorphic decryption. It also proposed a theoretical model to examine the error behavior of secure inference and presented parameters that can achieve smaller ciphertext size.…”

Section: Cryptography Homomorphicmentioning

confidence: 99%

Security in Approximate Computing and Approximate Computing for Security: Challenges and Opportunities

Liu

O’Neill

et al. 2020

Proc. IEEE

View full text Add to dashboard Cite

Approximate computing, an advanced computational technique which returns inaccurate but acceptable results instead of exact results, has emerged as a new preferable paradigm over traditional computing architectures for energy efficient system designs. It is crucial for nanoscale integrated circuits (ICs) to achieve high speed and low power, where some intrinsic errors are acceptable, such as (deep-) machine learning, image processing, communication and other error-tolerant and cognitive applications. However, approximate computing also introduces security vulnerabilities mainly due to the uncertainty and unpredictability of intrinsic errors during approximate execution which may be indistinguishable using malicious modification of the accurate result. On the other hand, interestingly, approximate computing can also provide new approaches for security. Existing literature in approximate computing covers threat models, countermeasures, and evaluations, but lacks a framework for analysis and comparison. In this paper, we provide a classification of the state of the art in this research field, including threat models in approximate computing and promising security approaches using approximate computing. Index Terms-Approximate computing, hardware security, cryptography I. INTRODUCTION N the last decade, various advanced computing systems, including supercomputers, ubiquitous computing centers, and servers, have been developed and widely deployed. Unfortunately, Moore's law is approaching its limitation [1], and conventional computing techniques are not able to provide higher computing performance under the restriction of power consumption. Therefore, new nanoscale computing paradigms are urgently required for low power and high performance computing systems. Hence, appropriate reduction of the computational accuracy could effectively improve the performance of computing systems without sacrificing functionality and perception.

show abstract

Section: Cryptography Homomorphicmentioning

confidence: 99%

Security in Approximate Computing and Approximate Computing for Security: Challenges and Opportunities

Liu

O’Neill

et al. 2020

Proc. IEEE

View full text Add to dashboard Cite

show abstract

“…Observe that different from [3], we do not need an expensive simulation to ensure an asymptotically small (e.g., 2 −40 ) decryption failure probability, since NN-based SI mispredicts much more often than 2 −40 . In most cases, a 0.1% accuracy degradation is not noticeable for practical CNN applications.…”

Section: Generating a Valid Ciphertext Modulusmentioning

confidence: 99%

“…• Optimizing HE Parameters: While existing works have already treated the instantiation of HE parameters as a design problem and proposed some solutions [3], we point out that these solutions are not adequate. In particular, we identify an optimization dilemma in learning with errors (LWE) based HE parameter instantiation, and observe that this optimization problem is (computationally) rather difficult to solve, especially for NAS-based optimization with fast turnaround time.…”

Section: Introductionmentioning

confidence: 99%

“…Recent advances in the secure machine learning field have suggested the possibility of formulating the secure protocols as a design automation problem. For example, in [3], authors proposed a framework that automatically instantiate parameters for homomorphic encryption (HE) schemes. Likewise, a line of research efforts [7,14,16] have explored how to optimize HE parameters and packing capabilities to improve the efficiency of secure computations, especially for neural network based protocols.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

NASS: Optimizing Secure Inference via Neural Architecture Search

Bian,

Jiang,

et al. 2020

Preprint

Self Cite

View full text Add to dashboard Cite

Due to increasing privacy concerns, neural network (NN) based secure inference (SI) schemes that simultaneously hide the client inputs and server models attract major research interests. While existing works focused on developing secure protocols for NN-based SI, in this work, we take a different approach. We propose NASS, an integrated framework to search for tailored NN architectures designed specifically for SI. In particular, we propose to model cryptographic protocols as design elements with associated reward functions. The characterized models are then adopted in a joint optimization with predicted hyperparameters in identifying the best NN architectures that balance prediction accuracy and execution efficiency. In the experiment, it is demonstrated that we can achieve the best of both worlds by using NASS, where the prediction accuracy can be improved from 81.6% to 84.6%, while the inference runtime is reduced by 2x and communication bandwidth by 1.9x on the CIFAR-10 dataset.

show abstract

“…Privacy is especially important, when clients upload their sensitive information, e.g., healthcare records and financial data, to cloud servers. Recent works [1,2,3,4,5] create Hybrid Privacy-Preserving Neural Networks (HPPNNs) to achieve high inference accuracy using a combination of Homomorphic Encryption (HE) and Garbled Circuit (GC). Particularly, DELPHI [5] obtains the state-of-the-art inference latency and accuracy through implementing linear layers by HE, and computing activation layers by GC.…”

Section: Introductionmentioning

confidence: 99%

AutoPrivacy: Automated Layer-wise Parameter Selection for Secure Neural Network Inference

Lou,

Bian,

Jiang

2020

Preprint

Self Cite

View full text Add to dashboard Cite

Hybrid Privacy-Preserving Neural Network (HPPNN) implementing linear layers by Homomorphic Encryption (HE) and nonlinear layers by Garbled Circuit (GC) is one of the most promising secure solutions to emerging Machine Learning as a Service (MLaaS). Unfortunately, a HPPNN suffers from long inference latency, e.g., ∼ 100 seconds per image, which makes MLaaS unsatisfactory. Because HE-based linear layers of a HPPNN cost 93% inference latency, it is critical to select a set of HE parameters to minimize computational overhead of linear layers. Prior HPPNNs over-pessimistically select huge HE parameters to maintain large noise budgets, since they use the same set of HE parameters for an entire network and ignore the error tolerance capability of a network. In this paper, for fast and accurate secure neural network inference, we propose an automated layer-wise parameter selector, AutoPrivacy, that leverages deep reinforcement learning to automatically determine a set of HE parameters for each linear layer in a HPPNN. The learning-based HE parameter selection policy outperforms conventional rule-based HE parameter selection policy. Compared to prior HPPNNs, AutoPrivacy-optimized HPPNNs reduce inference latency by 53% ∼ 70% with negligible loss of accuracy.

show abstract

DArL: Dynamic Parameter Adjustment for LWE-based Secure Inference

Cited by 6 publications

References 21 publications

Security in Approximate Computing and Approximate Computing for Security: Challenges and Opportunities

Security in Approximate Computing and Approximate Computing for Security: Challenges and Opportunities

NASS: Optimizing Secure Inference via Neural Architecture Search

AutoPrivacy: Automated Layer-wise Parameter Selection for Secure Neural Network Inference

Contact Info

Product

Resources

About