Federated learning (FL) is a distributed Machine Learning (ML) framework that is capable of training a new global model by aggregating clients' locally trained models without sharing users' original data. Fede...
详细信息
ISBN:
(纸本)9789819608041;9789819608058
Federated learning (FL) is a distributed Machine Learning (ML) framework that is capable of training a new global model by aggregating clients' locally trained models without sharing users' original data. Federated learning as a service (FLaaS) offers a privacy-preserving approach for training machine learning models on devices with various computational resources. Most proposed FL-based methods train the same model in all client devices regardless of their computational resources. However, in practical Internet of Things (IoT) scenarios, IoT devices with limited computational resources may not be capable of training models that client devices with greater hardware performance hosted. Most of the existing FL frameworks that aim to solve the problem of aggregating heterogeneous models are designed for Independent and Identical distributed (IID) data, which may make it hard to reach the target algorithm performance when encountering non-IID scenarios. To address these problems in hierarchical networks, in this paper, we propose a heterogeneous aggregation framework for hierarchical edge systems called HAF-Edge. In our proposed framework, we introduce a communication-efficient model aggregation method designed for FL systems with two-level model aggregations running at the edge and cloud levels. This approach enhances the convergence rate of the global model by leveraging selective knowledge transfer during the aggregation of heterogeneous models. To the best of our knowledge, this work is pioneering in addressing the problem of aggregating heterogeneous models within hierarchical FL systems spanning IoT, edge, and cloud environments. We conducted extensive experiments to validate the performance of our proposed method. The evaluation results demonstrate that HAF-Edge significantly outperforms state-of-the-art methods.
Object detection is one of the most fundamental problems in computer vision, and image sensors are commonly used for this. In this paper, we present the impact of image sensor faults on pruned neural networks for obje...
详细信息
Natural Language Processing (NLP) models are one of the most promising topics nowadays. Applications like ChatGPT uncover the power of such models and their broad applications. However, developing such models requires...
详细信息
This paper explores significant advances in machine learning (ML) in the field of natural language processing (NLP), with an emphasis on transformative innovations such as transformer models and large language models ...
详细信息
Currently, semi-supervised learning methods have become popular in medical image processing tasks, this paper proposes a medical image classification method based on a semi-supervised pseudo-label augmented generative...
详细信息
Action recognition is a prerequisite for many applications in laparoscopic video analysis including but not limited to surgical training, operation room planning, follow-up surgery preparation, post-operative surgical...
详细信息
ISBN:
(纸本)9798350312249
Action recognition is a prerequisite for many applications in laparoscopic video analysis including but not limited to surgical training, operation room planning, follow-up surgery preparation, post-operative surgical assessment, and surgical outcome estimation. However, automatic action recognition in laparoscopic surgeries involves numerous challenges such as (I) cross-action and intra-action duration variation, (II) relevant content distortion due to smoke, blood accumulation, fast camera motions, organ movements, object occlusion, and (III) surgical scene variations due to different illuminations and viewpoints. Besides, action annotations in laparoscopy surgeries are limited and expensive due to requiring expert knowledge. In this study, we design and evaluate a CNN-RNN architecture as well as a customized training-inference framework to deal with the mentioned challenges in laparoscopic surgery action recognition. Using stacked recurrent layers, our proposed network takes advantage of inter-frame dependencies to negate the negative effect of content distortion and variation in action recognition. Furthermore, our proposed frame sampling strategy effectively manages the duration variations in surgical actions to enable action recognition with high temporal resolution. Our extensive experiments confirm the superiority of our proposed method in action recognition compared to static CNNs.
Phishing emails are sent by thousands of phishers who attempt to trick a user into believing that the email is from someone they know or trust in order for them to take some sort of action upon opening it. Despite the...
详细信息
The rise of autonomous and connected vehicles in today's transportation systems has highlighted the importance of exchanging multimedia data efficiently and securely. Traditional centralized methods of storing and...
详细信息
The proceedings contain 18 papers. The topics discussed include: Optimal RTGC Non-linear Control System Based on Sliding Mode Controller;Scalar Modules with Logarithmic Lyapunov Functions for Robust Positive systems E...
ISBN:
(纸本)9784907764814
The proceedings contain 18 papers. The topics discussed include: Optimal RTGC Non-linear Control System Based on Sliding Mode Controller;Scalar Modules with Logarithmic Lyapunov Functions for Robust Positive systems Exhibiting Interior Equilibria;Broadband Modeling of Power Distribution networks;A Non-contact Translational and Rotational Force Feedback Device using Rotational Jet Propellers;Incentivizing Control Design for Blockchain-Enabled distributed Optimization in Electricity Market systems;Interception of Multiple Drone Targets by Heterogeneous Chasers using Heuristic Task Allocations with DQN-GNN guidance model;characterizing the dynamic behavior of a both-sides retrodirective system for the control of microwave wireless power transfer;and data-driven optimal control with data loss.
Various methods have been developed for automated and semi-automated architecture generation in the computer aided ship design processes. The question remains as to how this can speed up the design process without los...
详细信息
Various methods have been developed for automated and semi-automated architecture generation in the computer aided ship design processes. The question remains as to how this can speed up the design process without losing the requirement elucidation intent for concept phase. This paper presents a novel approach with a software toolset to develop design and analysis approaches to early stage ship design and provide a sketching tool. This was done by enhancing the user interface and experience of the UCL Network Block Approach to achieve a "thinking sketch" in a way that is "quick" and "fluid" enough to promote inventive and creative sketching comparable to hand sketching. The UCL Network Block Approach draws on the UCL Design Building Block (DBB) approach and uses network methods applied to the synthesis of distributed ship service systems (DS3) and computer Aided Ship Design (CASD) to expand DS3 definition in early stage ship design. The UCL originated inside-out/DBB approach to sketch driven synthesis has been made translatable to both DBB ship descriptions and ensuring early stage naval architectural "balance". The proposed approach has been used for the first time successfully to not only carry out a rapid sketching exercise for a naval ship design but also enable quick preliminary analysis of a set of DS3 networks.
暂无评论