EdgeEye

Liu, Peng; Qi, Bozhao; Banerjee, Suman

doi:10.1145/3213344.3213345

Cited by 67 publications

(5 citation statements)

References 7 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Split Federated Learning Scheme Detail The idea of splitting the model into a client-side model and a server-side model was first proposed in (Kang et al 2017;Teerapittayanon, McDanel, and Kung 2017;Liu, Qi, and Banerjee 2018) for inference tasks and extended by (Thapa et al 2020) into Split Federated Learning (SFL), a collaborative learning scheme suitable for resource-constrained devices. In SFL, clients perform forward propagation locally till the last layer of client-side model, sending the intermediate activation with label information to the server.…”

Section: Related Workmentioning

confidence: 99%

EMGAN: Early-Mix-GAN on Extracting Server-Side Model in Split Federated Learning

Li,

Chen,

Yang

et al. 2024

AAAI

View full text Add to dashboard Cite

Split Federated Learning (SFL) is an emerging edge-friendly version of Federated Learning (FL), where clients process a small portion of the entire model. While SFL was considered to be resistant to Model Extraction Attack (MEA) by design, a recent work shows it is not necessarily the case. In general, gradient-based MEAs are not effective on a target model that is changing, as is the case in training-from-scratch applications. In this work, we propose a strong MEA during the SFL training phase. The proposed Early-Mix-GAN (EMGAN) attack effectively exploits gradient queries regardless of data assumptions. EMGAN adopts three key components to address the problem of inconsistent gradients. Specifically, it employs (i) Early-learner approach for better adaptability, (ii) Multi-GAN approach to introduce randomness in generator training to mitigate mode collapse, and (iii) ProperMix to effectively augment the limited amount of synthetic data for a better approximation of the target domain data distribution. EMGAN achieves excellent results in extracting server-side models. With only 50 training samples, EMGAN successfully extracts a 5-layer server-side model of VGG-11 on CIFAR-10, with 7% less accuracy than the target model. With zero training data, the extracted model achieves 81.3% accuracy, which is significantly better than the 45.5% accuracy of the model extracted by the SoTA method. The code is available at "https://github.com/zlijingtao/SFL-MEA".

show abstract

Section: Related Workmentioning

confidence: 99%

EMGAN: Early-Mix-GAN on Extracting Server-Side Model in Split Federated Learning

Li,

Chen,

Yang

et al. 2024

AAAI

View full text Add to dashboard Cite

show abstract

“…The key idea for model-split learning schemes is to split the model so that part of it is processed in the client and the rest is offloaded to the server. This idea was first proposed in [13,17,26] for inference tasks and extended by [9] for split learning, a collaborative multi-client neural network training. However, the round-robin design in [9] need clients to learn sequentially and thus required long training time.…”

Section: Related Work 21 Model-split Learning Schemesmentioning

confidence: 99%

Model Extraction Attacks on Split Federated Learning

Li¹,

Rakin²,

Chen³

et al. 2023

Preprint

View full text Add to dashboard Cite

Federated Learning (FL) is a popular collaborative learning scheme involving multiple clients and a server. FL focuses on protecting clients' data but turns out to be highly vulnerable to Intellectual Property (IP) threats. Since FL periodically collects and distributes the model parameters, a free-rider can download the latest model and thus steal model IP. Split Federated Learning (SFL), a recent variant of FL that supports training with resource-constrained clients, splits the model into two, giving one part of the model to clients (client-side model), and the remaining part to the server (server-side model). Thus SFL prevents model leakage by design. Moreover, by blocking prediction queries, it can be made resistant to advanced IP threats such as traditional Model Extraction (ME) attacks. While SFL is better than FL in terms of providing IP protection, it is still vulnerable. In this paper, we expose the vulnerability of SFL and show how malicious clients can launch ME attacks by querying the gradient information from the server side. We propose five variants of ME attack which differs in the gradient usage as well as in the data assumptions. We show that under practical cases, the proposed ME attacks work exceptionally well for SFL. For instance, when the server-side model has five layers, our proposed ME attack can achieve over 90% accuracy with less than 2% accuracy degradation with VGG-11 on CIFAR-10.

show abstract

“…However, their platform design resembles a general-purpose event-driven middleware, without any specific analytics support or runtime optimizations for video processing, unlike us. Edge-Eye [41] efficiently deploys DNN models on the edge, using a JavaScript API for users to specify their parameters. It offers performance optimizations for DNNs, but does not consider distributed systems issues, such as batching, dropping and network variability.…”

Section: Related Work 61 Video Surveillance Systemsmentioning

confidence: 99%

A Scalable Platform for Distributed Object Tracking Across a Many-Camera Network

Khochare

Krishnan

Simmhan

2021

IEEE Trans. Parallel Distrib. Syst.

View full text Add to dashboard Cite

Advances in deep neural networks (DNN) and computer vision (CV) algorithms have made it feasible to extract meaningful insights from large-scale deployments of urban cameras. Tracking an object of interest across the camera network in near real-time is a canonical problem. However, current tracking platforms have two key limitations: 1) They are monolithic, proprietary and lack the ability to rapidly incorporate sophisticated tracking models, and 2) They are less responsive to dynamism across wide-area computing resources that include edge, fog, and cloud abstractions. We address these gaps using Anveshak, a runtime platform for composing and coordinating distributed tracking applications. It provides a domain-specific dataflow programming model to intuitively compose a tracking application, supporting contemporary CV advances like query fusion and re-identification, and enabling dynamic scoping of the camera network's search space to avoid wasted computation. We also offer tunable batching and data-dropping strategies for dataflow blocks deployed on distributed resources to respond to network and compute variability. These balance the tracking accuracy, its real-time performance, and the active camera-set size. We illustrate the concise expressiveness of the programming model for four tracking applications. Our detailed experiments for a network of 1000 camera-feeds on modest resources exhibit the tunable scalability, performance, and quality trade-offs enabled by our dynamic tracking, batching, and dropping strategies.

show abstract

EdgeEye

Cited by 67 publications

References 7 publications

EMGAN: Early-Mix-GAN on Extracting Server-Side Model in Split Federated Learning

EMGAN: Early-Mix-GAN on Extracting Server-Side Model in Split Federated Learning

Model Extraction Attacks on Split Federated Learning

A Scalable Platform for Distributed Object Tracking Across a Many-Camera Network

Contact Info

Product

Resources

About