An Accelerated Doubly Stochastic Gradient Method with Faster Explicit Model Identification

Bao, Runxue; Gu, Bin; Huang, Heng

doi:10.1145/3511808.3557234

Cited by 3 publications

(3 citation statements)

References 23 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…51 Furthermore, there are studies on GAP safe screening for optimization algorithm based on stochastic gradient descent. 52 , 53 …”

Section: Resultsmentioning

confidence: 99%

See 1 more Smart Citation

Efficient model selection for predictive pattern mining model by safe pattern pruning

Yoshida,

Hanada,

Nakagawa

et al. 2023

Patterns

View full text Add to dashboard Cite

“…51 Furthermore, there are studies on GAP safe screening for optimization algorithm based on stochastic gradient descent. 52 , 53 …”

Section: Resultsmentioning

confidence: 99%

“…Several studies exist in the paradigm of stochastic gradient method. 52 , 53 Although it is not clear that our method can be directly applied to them, this topic is worth discussing in future work. Additionally, the sphere refinement technique 51 can further accelerate SPP more in some loss functions.…”

Section: Discussionmentioning

confidence: 99%

Efficient model selection for predictive pattern mining model by safe pattern pruning

Yoshida,

Hanada,

Nakagawa

et al. 2023

Patterns

View full text Add to dashboard Cite

“…The challenges of MLaaS comes from several folds: inference latency and privacy. To accelerate the MLaaS training and inference application, accelerated gradient sparsification [12], [13] and model compression methods [14]- [22] are proposed. On the other side, a major limitation of MLaaS is the requirement for clients to reveal raw input data to the service provider, which may compromise the privacy of users.…”

Section: Introductionmentioning

confidence: 99%

RRNet: Towards ReLU-Reduced Neural Network for Two-party Computation Based Private Inference

Peng¹,

Zhou²,

Luo³

et al. 2023

Preprint

View full text Add to dashboard Cite

The proliferation of deep learning (DL) has led to the emergence of privacy and security concerns. To address these issues, secure Two-party computation (2PC) has been proposed as a means of enabling privacy-preserving DL computation. However, in practice, 2PC methods often incur high computation and communication overhead, which can impede their use in largescale systems. To address this challenge, we introduce RRNet, a systematic framework that aims to jointly reduce the overhead of MPC comparison protocols and accelerate computation through hardware acceleration. Our approach integrates the hardware latency of cryptographic building blocks into the DNN loss function, resulting in improved energy efficiency, accuracy, and security guarantees. Furthermore, we propose a cryptographic hardware scheduler and corresponding performance model for Field Programmable Gate Arrays (FPGAs) to further enhance the efficiency of our framework. Experiments show RRNet achieved a much higher ReLU reduction performance than all SOTA works on CIFAR-10 dataset.

show abstract

Self-Adaptive Perturbation Radii for Adversarial Training

Shi

Zhang

2023

Proceedings of the 29th ACM SIGKDD Conference on Knowledge Discovery and Data Mining

View full text Add to dashboard Cite

An Accelerated Doubly Stochastic Gradient Method with Faster Explicit Model Identification

Cited by 3 publications

References 23 publications

Efficient model selection for predictive pattern mining model by safe pattern pruning

Efficient model selection for predictive pattern mining model by safe pattern pruning

RRNet: Towards ReLU-Reduced Neural Network for Two-party Computation Based Private Inference

Self-Adaptive Perturbation Radii for Adversarial Training

Contact Info

Product

Resources

About