Federated transfer learning for disaster classification in social computing networks

Zhang, Zehui; He, Ningxin; Li, Donagyu; Gao, Hang; Gao, Tiegang; Zhou, Chuan

doi:10.1016/j.jnlssr.2021.10.007

Cited by 15 publications

(7 citation statements)

References 31 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Thirdly, our proposed method proved the FTL framework has compatibility and feasibility in a more realistic and complete production environment. In Zhang et al ( 2021 , 2022 ), this machinery fault FTL has achieved about 95% accuracy with the CWRU dataset which is relatively clean with limited noise, and about 90% accuracy with the crack dataset with a test rig that consists of two simple bearings. In contrast, even though the accuracy of our proposed method is 81%, its data acquisition was from a real-life PCA production line with more complexity and interference.…”

Section: Discussionmentioning

confidence: 99%

“…After that TL allows that global model to share knowledge between tasks from different but related domains (Saha & Ahmad, 2021 ). (2) TL is a part of FL, that TL is adopted as an approach to establish a personalized model, which can decrease the computation and communication costs in the FL system (Wang et al, 2022 ; Zhang et al, 2022 ). (3) FTL allows knowledge to be shared with data privacy-preservation of the clients and enables knowledge to be transferred from the source domain to the target domain in a data federation (Liu et al, 2020a ).…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Federated transfer learning for auxiliary classifier generative adversarial networks: framework and industrial application

et al. 2023

View full text Add to dashboard Cite

Machine learning with considering data privacy-preservation and personalized models has received attentions, especially in the manufacturing field. The data often exist in the form of isolated islands and cannot be shared because of data privacy in real industrial scenarios. It is difficult to gather the data to train a personalized model without compromising data privacy. To address this issue, we proposed a Federated Transfer Learning framework based on Auxiliary Classifier Generative Adversarial Networks named ACGAN-FTL. In the framework, Federated Learning (FL) trains a global model on decentralized datasets of the clients with data privacy-preservation and Transfer Learning (TL) transfers the knowledge from the global model to a personalized model with a relatively small data volume. ACGAN acts as a data bridge to connect FL and TL by generating similar probability distribution data of clients since the client datasets in FL cannot be directly used in TL for data privacy-preservation. A real industrial scenario of pre-baked carbon anode quality prediction is applied to verify the performance of the proposed framework. The results show that ACGAN-FTL can not only obtain acceptable performance on 0.81 accuracy, 0.86 precision, 0.74 recall, and 0.79 F1 but also ensure data privacy-preservation in the whole learning process. Compared to the baseline method without FL and TL, the former metrics have increased by 13%, 11%, 16%, and 15% respectively. The experiments verify that the performance of the proposed ACGAN-FTL framework fulfills the requirements of industrial scenarios.

show abstract

Section: Discussionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Federated transfer learning for auxiliary classifier generative adversarial networks: framework and industrial application

et al. 2023

View full text Add to dashboard Cite

show abstract

“…FL could prove a more effective training model in disaster scenarios due to massive irrelevant data and a high flow of scalable IoT networks while in disasters. The privacy-preserved federated transfer learning (FedTL) approach could also provide and verify a real disaster image dataset collected from distributed social computing nodes (97). The new-developed FL model will give both AI and humans (e.g., first responders) a sophisticated training course and achieve better performance for the medical image learning model (98).…”

Section: Machine Learning Applied In Dm Educationmentioning

confidence: 99%

The development of new remote technologies in disaster medicine education: A scoping review

Kao

Chien

Wang

et al. 2023

Front. Public Health

View full text Add to dashboard Cite

BackgroundRemote teaching and online learning have significantly changed the responsiveness and accessibility after the COVID-19 pandemic. Disaster medicine (DM) has recently gained prominence as a critical issue due to the high frequency of worldwide disasters, especially in 2021. The new artificial intelligence (AI)-enhanced technologies and concepts have recently progressed in DM education.ObjectivesThe aim of this article is to familiarize the reader with the remote technologies that have been developed and used in DM education over the past 20 years.Literature scoping reviewsMobile edge computing (MEC), unmanned aerial vehicles (UAVs)/drones, deep learning (DL), and visual reality stimulation, e.g., head-mounted display (HMD), are selected as promising and inspiring designs in DM education.MethodsWe performed a comprehensive review of the literature on the remote technologies applied in DM pedagogy for medical, nursing, and social work, as well as other health discipline students, e.g., paramedics. Databases including PubMed (MEDLINE), ISI Web of Science (WOS), EBSCO (EBSCO Essentials), Embase (EMB), and Scopus were used. The sourced results were recorded in a Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) flowchart and followed in accordance with the PRISMA extension Scoping Review checklist. We included peer-reviewed articles, Epubs (electronic publications such as databases), and proceedings written in English. VOSviewer for related keywords extracted from review articles presented as a tabular summary to demonstrate their occurrence and connections among these DM education articles from 2000 to 2022.ResultsA total of 1,080 research articles on remote technologies in DM were initially reviewed. After exclusion, 64 articles were included in our review. Emergency remote teaching/learning education, remote learning, online learning/teaching, and blended learning are the most frequently used keywords. As new remote technologies used in emergencies become more advanced, DM pedagogy is facing more complex problems.DiscussionsArtificial intelligence-enhanced remote technologies promote learning incentives for medical undergraduate students or graduate professionals, but the efficacy of learning quality remains uncertain. More blended AI-modulating pedagogies in DM education could be increasingly important in the future. More sophisticated evaluation and assessment are needed to implement carefully considered designs for effective DM education.

show abstract

“…FTL, which is inspired by the transfer learning model, aims to provide ML approaches in cases where entities suffer from insufficient data samples. For instance, some data are available from a domain (electric bus) that can be used in a prediction model in another EV domain with a limited amount of available data [70].…”

Section: Federated Learningmentioning

confidence: 99%

A Review on Intelligent Energy Management Systems for Future Electric Vehicle Transportation

Teimoori

Yassine

2022

Sustainability

View full text Add to dashboard Cite

Over the last few years, Electric Vehicles (EVs) have been gaining interest as a result of their ability to reduce vehicle emissions. Developing an intelligent system to manage EVs charging demands is one of the fundamental aspects of this technology to better adapt for all-purpose transportation utilization. It is necessary for EVs to be connected to the Smart Grid (SG) to communicate with charging stations and other energy resources in order to control charging schedules, while Artificial Intelligent (AI) techniques can be beneficial for improving the system, they can also raise security and privacy threats. In recent years, privacy preservation methodologies have been introduced to ensure data security. Federated Learning (FL) and blockchain technology are two emerging strategies to address information protection concerns. Therefore, in this article, a comprehensive literature review is proposed to analyze existing EVs energy management challenges and solutions and present potential future research directions for EVs charging/discharging coordination applications.

show abstract

Federated transfer learning for disaster classification in social computing networks

Cited by 15 publications

References 31 publications

Federated transfer learning for auxiliary classifier generative adversarial networks: framework and industrial application

Federated transfer learning for auxiliary classifier generative adversarial networks: framework and industrial application

The development of new remote technologies in disaster medicine education: A scoping review

A Review on Intelligent Energy Management Systems for Future Electric Vehicle Transportation

Contact Info

Product

Resources

About