• Proceedings of the Korea Information Processing Society Conference
• /
• 2023.11a
• /
• pp.140-142
• /
• 2023

Federated learning (FL) is a new paradigm in machine learning (ML) that enables multiple devices to collaboratively train a shared ML model without sharing their local data. FL is well-suited for applications where data is sensitive or difficult to transmit in large volumes, or where collaborative learning is required. The Internet of Underwater Things (IoUT) is a network of underwater devices that collect and exchange data. This data can be used for a variety of applications, such as monitoring water quality, detecting marine life, and tracking underwater vehicles. However, the harsh underwater environment makes it difficult to collect and transmit data in large volumes. FL can address these challenges by enabling devices to train a shared ML model without having to transmit their data to a central server. This can help to protect the privacy of the data and improve the efficiency of training. In this view, this paper provides a brief overview of Fed-IoUT, highlighting its various applications, challenges, and opportunities.

• Journal of the Korea Institute of Information Security & Cryptology
• /
• v.34 no.1
• /
• pp.143-156
• /
• 2024

Federated Learning (FL) has emerged as a potent methodology for decentralized model training across multiple collaborators, eliminating the need for data sharing. Although FL is lauded for its capacity to preserve data privacy, it is not impervious to various types of privacy attacks. Differential Privacy (DP), recognized as the golden standard in privacy-preservation techniques, is widely employed to counteract these vulnerabilities. This paper makes a specific contribution by applying an existing, task-specific adaptive DP mechanism to the FL environment. Our comprehensive analysis evaluates the impact of this mechanism on the performance of a shared global model, with particular attention to varying data distribution and partitioning schemes. This study deepens the understanding of the complex interplay between privacy and utility in FL, providing a validated methodology for securing data without compromising performance.

• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2021.05a
• /
• pp.2-4
• /
• 2021

In a distributed machine learning system, sharing gradients was considered safe because it did not share original training data. However, recent studies found that malicious attacker could completely restore the original training data from shared gradients. Gradient Leakage Attack is a technique that restoring original training data by exploiting theses vulnerability. In this study, we present the image transformation method based on Discrete Cosine Transform to defend against the Gradient Leakage Attack on the federated learning setting, which training in local devices and sharing gradients to the server. Experiment shows that our image transformation method cannot be completely restored the original data from Gradient Leakage Attack.

• Smart Media Journal
• /
• v.13 no.4
• /
• pp.33-48
• /
• 2024

With the advent of 5G, characterized by Enhanced Mobile Broadband (eMBB), Ultra-Reliable Low Latency Communications (URLLC), and Massive Machine Type Communications (mMTC), efficient network management and service provision are becoming increasingly critical. This paper proposes a novel approach to address key challenges of 5G networks, namely ultra-high speed, ultra-low latency, and ultra-reliability, while dynamically optimizing network slicing and resource allocation using machine learning (ML) and deep learning (DL) techniques. The proposed methodology utilizes prediction models for network traffic and resource allocation, and employs Federated Learning (FL) techniques to simultaneously optimize network bandwidth, latency, and enhance privacy and security. Specifically, this paper extensively covers the implementation methods of various algorithms and models such as Random Forest and LSTM, thereby presenting methodologies for the automation and intelligence of 5G network operations. Finally, the performance enhancement effects achievable by applying ML and DL to 5G networks are validated through performance evaluation and analysis, and solutions for network slicing and resource management optimization are proposed for various industrial applications.

• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2022.10a
• /
• pp.597-599
• /
• 2022

When IoT device performs on-device AI, the device is required to use various AI models selectively according to target service and surrounding environment. Also, AI model can be updated by additional training such as federated learning or adapting the improved technique. Hence, for successful on-device AI, IoT device should acquire various AI models selectively or update previous AI model to new one. In this paper, we propose AI model repository to tackle this issue. The repository supports AI model registration, searching, management, and deployment along with dashboard for practical usage. We implemented it using Node.js and Vue.js to verify it works well.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2021.05a
• /
• pp.198-201
• /
• 2021

무인 항공기(UAV, Unmanned Aerial Vehicles)는 높은 기동성을 가지며 설치 비용이 저렴하다는 이점이 있어 홍수, 지진 등의 재난 재해 감시 시스템에 이용되고 있다. 재난 재해 감시 시스템에서 UAV는 지상에 위치한 사물인터넷(IoT, Internet of Things) 기기로부터 데이터를 수집하는 임무를 수행하기 위해 계획된 항로를 따라 비행한다. 이때 UAV가 정상 경로로 비행하기 위해서는 실시간으로 GPS 위치 확인이 가능해야 한다. 만일 UAV가 계산한 현재 위치의 GPS 정보가 잘못될 경우 비행경로에 대한 통제권을 상실하여 임무 수행을 완료하지 못하는 결과가 초래될 수 있다는 취약점이 존재한다. 이러한 취약점으로 인해 UAV는 공격자가 악의적으로 거짓 GPS 위치 신호를 전송하는GPS 스푸핑(Spoofing) 공격에 쉽게 노출된다. 본 논문에서는 신뢰할 수 있는 시스템을 구축하기 위해 지상에 위치한 기기가 송신하는 신호의 세기와 GPS 정보를 이용하여 UAV에 GPS 스푸핑 공격 여부를 탐지하고 공격당한 UAV가 경로를 이탈하지 않도록 대응하기 위해 연합학습(Federated Learning)을 이용하는 방안을 제안한다.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.18 no.7
• /
• pp.1774-1794
• /
• 2024

In the field of environmental sensing, it is necessary to develop radio planning techniques for the next generation Internet of Things (IoT) networks over mixed terrains. Such techniques are needed for smart remote monitoring of utility supplies, with links situated close to but out of range of cellular networks. In this paper, a three-dimension (3-D) geometric optimization algorithm is proposed, considering the positions of edge IoT devices and antenna coupling factors. Firstly, a multi-level single linkage (MLSL) iteration method, based on geometric objectives, is derived to evaluate the data rates over ISM 915 MHz channels, utilizing optimized power-distance profiles of continuous waves. Subsequently, a federated learning (FL) data selection algorithm is designed based on the 3-D geometric positions. Finally, a measurement example is taken in a meadow biome of the Mexican Colima district, which is prone to fluvial floods. The empirical path loss model has been enhanced, demonstrating the accuracy of the proposed optimization algorithm as well as the possibility of further prediction work.

• The Journal of Bigdata
• /
• v.8 no.2
• /
• pp.125-131
• /
• 2023

Globally, drug side effects rank among the top causes of death. To effectively respond to these adverse drug reactions, a shift towards an active real-time monitoring system, along with the standardization and quality improvement of data, is necessary. Integrating individual institutional data and utilizing large-scale data to enhance the accuracy of drug side effect predictions is critical. However, data sharing between institutions poses privacy concerns and involves varying data standards. To address this issue, our research adopts a federated learning approach, where data is not shared directly in compliance with privacy regulations, but rather the results of the model's learning are shared. We employ the Common Data Model (CDM) to standardize different data formats, ensuring accuracy and consistency of data. Additionally, we propose a drug monitoring system that enhances security and scalability management through a cloud-based federated learning environment. This system allows for effective monitoring and prediction of drug side effects while protecting the privacy of data shared between hospitals. The goal is to reduce mortality due to drug side effects and cut medical costs, exploring various technical approaches and methodologies to achieve this.

• Journal of the Korea Society of Computer and Information
• /
• v.27 no.11
• /
• pp.13-18
• /
• 2022

In this paper, we proposed VW-FedAVG(Validation based Weighted FedAVG) which updates the global model by weighting according to performance verification from the models of each device participating in the training. The first method is designed to validate each local client model through validation dataset before updating the global model with a server side validation structure. The second is a client-side validation structure, which is designed in such a way that the validation data set is evenly distributed to each client and the global model is after validation. MNIST, CIFAR-10 is used, and the IID, Non-IID distribution for image classification obtained higher accuracy than previous studies.

• Convergence Security Journal
• /
• v.21 no.3
• /
• pp.49-56
• /
• 2021

Recently, artificial intelligence is regarded as an essential technology in our society. In particular, the invasion of privacy in artificial intelligence has become a serious problem in modern society. Split learning, proposed at MIT in 2019 for privacy protection, is a type of federated learning technique that does not share any raw data. In this study, we studied a safe and accurate segmentation learning model using known differential privacy to safely manage data. In addition, we trained SVHN and GTSRB on a split learning model to which 15 different types of differential privacy are applied, and checked whether the learning is stable. By conducting a learning data extraction attack, a differential privacy budget that prevents attacks is quantitatively derived through MSE.