Abstract:In order to extract knowledge from the large data collected by edge devices, traditional cloud based approach that requires data upload may not be feasible due to communication bandwidth limitation as well as privacy and security concerns of end users. To address these challenges, a novel privacy preserving edge computing framework is proposed in this paper for image classification. Specifically, autoencoder will be trained unsupervised at each edge device individually, then the obtained latent vectors will be transmitted to the edge server for the training of a classifier. This framework would reduce the communications overhead and protect the data of the end users. Comparing to federated learning, the training of the classifier in the proposed framework does not subject to the constraints of the edge devices, and the autoencoder can be trained independently at each edge device without any server involvement. Furthermore, the privacy of the end users' data is protected by transmitting latent vectors without additional cost of encryption. Experimental results provide insights on the image classification performance vs. various design parameters such as the data compression ratio of the autoencoder and the model complexity.

What problem does this paper attempt to address?

The problem that this paper attempts to solve is the communication bandwidth limitations, privacy and security issues faced by traditional cloud computing methods when performing image classification in edge computing systems. Specifically: 1. **Communication Bandwidth Limitations**: Uploading a large amount of image data may overwhelm the uplink, especially in the case of limited bandwidth. 2. **Privacy and Security Issues**: Users may be concerned about the privacy and security of their image data and be unwilling to upload the original images, which may contain sensitive information. In addition, there may be risks of eavesdropping, interception or other unauthorized access during the upload process. To solve these problems, the paper proposes a novel and efficient privacy - protecting edge computing framework. The core ideas of this framework are: - Independently train an autoencoder on each edge device to extract the key features of the image and compress these features into latent vectors. - Only transmit these latent vectors to the edge server instead of the original image, thereby reducing communication overhead and protecting the privacy of user data. - Use these latent vectors on the edge server to train the classifier without relying on the computing power of the edge device. In this way, this framework not only reduces communication overhead but also improves the security and privacy of data. Compared with federated learning, this framework has the following advantages: 1. **Unlimited Model Complexity**: The classifier can be trained on the cloud server, so it can be very deep and complex without being limited by the computing power and storage of the edge device. 2. **Independent Training**: The autoencoder can be independently trained on each edge device without the participation of the server. 3. **No Additional Encryption Cost**: By transmitting latent vectors instead of the original data, the "encryption" of data is naturally achieved without additional encryption steps. The experimental results show that, under an appropriate compression ratio, the image classification performance of this framework hardly decreases, and its performance under different design parameters also provides valuable insights.

Efficient Privacy Preserving Edge Computing Framework for Image Classification

A Distributed Privacy-Preserving Framework for Deep Learning with Edge-Cloud Computing.

Research on Image Classification Algorithm Based on Horizontal Federated Learning

RISE : Privacy preserved data analytics using Regularized Inference Specific autoEncoder

Edge-cloud Collaborative Learning with Federated and Centralized Features

Keep Your Data Locally: Federated-Learning-Based Data Privacy Preservation in Edge Computing

P-CA: Privacy-Preserving Convolutional Autoencoder-Based Edge–Cloud Collaborative Computing for Human Behavior Recognition

Edge-AI-Driven Framework with Efficient Mobile Network Design for Facial Expression Recognition

A privacy-preserving deep learning approach for face recognition with edge computing

Lightweight Privacy-Preserving Ensemble Classification for Face Recognition

PFLF: Privacy-Preserving Federated Learning Framework for Edge Computing

Edge-Enabled Distributed Deep Learning for 5G Privacy Protection

Edge Intelligence-based Privacy Protection Framework for IoT-based Smart Healthcare Systems

Construction Jobsite Image Classification Using an Edge Computing Framework

Edge-device Collaborative Computing for Multi-view Classification

Privacy-Preserving Federated Learning for Industrial Edge Computing via Hybrid Differential Privacy and Adaptive Compression

Survey: federated learning data security and privacy-preserving in edge-Internet of Things

Private Edge Computing Resource Allocation and Communication Optimization Based on Federated Learning

Edge-based secure image denoising scheme supporting flexible user authorization

Lightweight and Unobtrusive Privacy Preservation for Remote Inference Via Edge Data Obfuscation

Residual-INR: Communication Efficient On-Device Learning Using Implicit Neural Representation