Submit to Special Issue Submit Abstract to Special Issue Review for Sensors Propose a Special Issue

Journal Menu

Journal Browser

► Journal Browser

Emerging Machine Learning Techniques in Industrial Internet of Things

Print Special Issue Flyer
Special Issue Editors
Special Issue Information
Keywords
Published Papers

A special issue of Sensors (ISSN 1424-8220). This special issue belongs to the section "Internet of Things".

Deadline for manuscript submissions: 15 October 2024 | Viewed by 7720

Share This Special Issue

Special Issue Editors

Prof. Dr. Maozhen Li

E-Mail Website
Guest Editor

Department of Electronic and Computer Engineering, Brunel University London, Middlesex UB8 3PH, UK
Interests: high-performance computing (grid and cloud computing); big data analytics; intelligent systems
Special Issues, Collections and Topics in MDPI journals

Dr. Zhengwen Huang

E-Mail Website
Co-Guest Editor

College of Engineering, Design and Physical Sciences, Brunel University London, Uxbridge UB8 3PH, UK
Interests: evolutionary algorithm; gene expression programming; machine learning; data engineering for big data analytics in smart grid; HPC; smart manufacturing
Special Issues, Collections and Topics in MDPI journals

Dr. Yang Liu

E-Mail Website
Co-Guest Editor

College of Electrical Engineering, Sichuan University, Chengdu 610065, China
Interests: smart grid; high-performance computing

Prof. Dr. Mukhtaj Khan

E-Mail Website
Co-Guest Editor

Department of Information Technology, University of Haripur, Haripur, Pakistan
Interests: high performance computing; sensor networks

Special Issue Information

Dear Colleagues,

With the Industrial Internet of Things (IIoT) increasingly used in areas such as smart manufacturing, smart grids and smart cities, it has become imperative to develop new machine learning techniques to enable future IIOT systems to be more efficient in computation, secure in data acquisition and analysis, preserved in data privacy, and robust in decision making. For this purpose, this Special Issue wishes to solicit state-of-the-art research or works in progress on emerging machine learning techniques.

Potential topics include, but are not limited to, lightweight deep neural network models, neural network compression techniques, machine learning with knowledge engineering, data encryptions, data privacy preserving techniques, federated learning, knowledge distillation, and transfer learning. In addition, we welcome original research articles covering new IIOT applications, case studies, challenges and developments in IIoT, as well as theoretical works in making light-weight deep neural networks. We also intend to include research works on computing technologies in support of IIOT facilities such as fog computing, edge computing, computation offloading, and hybrid edge-fog-cloud computing in this Special Issue.

Prof. Dr. Maozhen Li
Dr. Zhengwen Huang
Dr. Yang Liu
Prof. Dr. Mukhtaj Khan
Guest Editors

Manuscript Submission Information

Manuscripts should be submitted online at www.mdpi.com by registering and logging in to this website. Once you are registered, click here to go to the submission form. Manuscripts can be submitted until the deadline. All submissions that pass pre-check are peer-reviewed. Accepted papers will be published continuously in the journal (as soon as accepted) and will be listed together on the special issue website. Research articles, review articles as well as short communications are invited. For planned papers, a title and short abstract (about 100 words) can be sent to the Editorial Office for announcement on this website.

Submitted manuscripts should not have been published previously, nor be under consideration for publication elsewhere (except conference proceedings papers). All manuscripts are thoroughly refereed through a single-blind peer-review process. A guide for authors and other relevant information for submission of manuscripts is available on the Instructions for Authors page. Sensors is an international peer-reviewed open access semimonthly journal published by MDPI.

Please visit the Instructions for Authors page before submitting a manuscript. The Article Processing Charge (APC) for publication in this open access journal is 2600 CHF (Swiss Francs). Submitted papers should be well formatted and use good English. Authors may use MDPI's English editing service prior to publication or during author revisions.

Keywords

industrial Internet of Things
computation efficient machine learning
federated learnin
knowledge distillation
edge computing
data privacy preserving
machine learning robustness

Published Papers (7 papers)

Download All Papers

Order results

Result details

Show export options Show export options

Select all

Export citation of selected articles as:

Research

22 pages, 1682 KiB

Open AccessArticle

Modeling Structured Dependency Tree with Graph Convolutional Networks for Aspect-Level Sentiment Classification

by Qin Zhao, Fuli Yang, Dongdong An and Jie Lian

Sensors 2024, 24(2), 418; https://doi.org/10.3390/s24020418 - 10 Jan 2024

Cited by 1 | Viewed by 632

Abstract

Aspect-based sentiment analysis is a fine-grained task where the key goal is to predict sentiment polarities of one or more aspects in a given sentence. Currently, graph neural network models built upon dependency trees are widely employed for aspect-based sentiment analysis tasks. However, most existing models still contain a large amount of noisy nodes that cannot precisely capture the contextual relationships between specific aspects. Meanwhile, most studies do not consider the connections between nodes without direct dependency edges but play critical roles in determining the sentiment polarity of an aspect. To address the aforementioned limitations, we propose a Structured Dependency Tree-based Graph Convolutional Network (SDTGCN) model. Specifically, we explore construction of a structured syntactic dependency graph by incorporating positional information, sentiment commonsense knowledge, part-of-speech tags, syntactic dependency distances, etc., to assign arbitrary edge weights between nodes. This enhances the connections between aspect nodes and pivotal words while weakening irrelevant node links, enabling the model to sufficiently express sentiment dependencies between specific aspects and contextual information. We utilize part-of-speech tags and dependency distances to discover relationships between pivotal nodes without direct dependencies. Finally, we aggregate node information by fully considering their importance to obtain precise aspect representations. Experimental results on five publicly available datasets demonstrate the superiority of our proposed model over state-of-the-art approaches; furthermore, the accuracy and F1-score show a significant improvement on the majority of datasets, with increases of 0.74, 0.37, 0.65, and 0.79, 0.75, 1.17, respectively. This series of enhancements highlights the effective progress made by the STDGCN model in enhancing sentiment classification performance. Full article

(This article belongs to the Special Issue Emerging Machine Learning Techniques in Industrial Internet of Things)

► Show Figures

Figure 1

30 pages, 5416 KiB

Open AccessArticle

Personalized Fair Split Learning for Resource-Constrained Internet of Things

by Haitian Chen, Xuebin Chen, Lulu Peng and Yuntian Bai

Sensors 2024, 24(1), 88; https://doi.org/10.3390/s24010088 - 23 Dec 2023

Cited by 1 | Viewed by 685

Abstract

With the flourishing development of the Internet of Things (IoT), federated learning has garnered significant attention as a distributed learning method aimed at preserving the privacy of participant data. However, certain IoT devices, such as sensors, face challenges in effectively employing conventional federated learning approaches due to limited computational and storage resources, which hinder their ability to train complex local models. Additionally, in IoT environments, devices often face problems of data heterogeneity and uneven benefit distribution between them. To address these challenges, a personalized and fair split learning framework is proposed for resource-constrained clients. This framework first adopts a U-shaped structure, dividing the model to enable resource-constrained clients to offload subsets of the foundational model to a central server while retaining personalized model subsets locally to meet the specific personalized requirements of different clients. Furthermore, to ensure fair benefit distribution, a model-aggregation method with optimized aggregation weights is used. This method reasonably allocates model-aggregation weights based on the contributions of clients, thereby achieving collaborative fairness. Experimental results demonstrate that, in three distinct data heterogeneity scenarios, employing personalized training through this framework exhibits higher accuracy compared to existing baseline methods. Simultaneously, the framework ensures collaborative fairness, fostering a more balanced and sustainable cooperation among IoT devices. Full article

(This article belongs to the Special Issue Emerging Machine Learning Techniques in Industrial Internet of Things)

► Show Figures

Figure 1

24 pages, 36717 KiB

Open AccessArticle

DaylilyNet: A Multi-Task Learning Method for Daylily Leaf Disease Detection

by Zishen Song, Dong Wang, Lizhong Xiao, Yongjian Zhu, Guogang Cao and Yuli Wang

Sensors 2023, 23(18), 7879; https://doi.org/10.3390/s23187879 - 14 Sep 2023

Viewed by 776

Abstract

Timely detection and management of daylily diseases are crucial to prevent yield reduction. However, detection models often struggle with handling the interference of complex backgrounds, leading to low accuracy, especially in detecting small targets. To address this problem, we propose DaylilyNet, an object detection algorithm that uses multi-task learning to optimize the detection process. By incorporating a semantic segmentation loss function, the model focuses its attention on diseased leaf regions, while a spatial global feature extractor enhances interactions between leaf and background areas. Additionally, a feature alignment module improves localization accuracy by mitigating feature misalignment. To investigate the impact of information loss on model detection performance, we created two datasets. One dataset, referred to as the ‘sliding window dataset’, was obtained by splitting the original-resolution images using a sliding window. The other dataset, known as the ‘non-sliding window dataset’, was obtained by downsampling the images. Experimental results in the ‘sliding window dataset’ and the ‘non-sliding window dataset’ demonstrate that DaylilyNet outperforms YOLOv5-L in mAP@0.5 by 5.2% and 4.0%, while reducing parameters and time cost. Compared to other models, our model maintains an advantage even in scenarios where there is missing information in the training dataset. Full article

(This article belongs to the Special Issue Emerging Machine Learning Techniques in Industrial Internet of Things)

► Show Figures

Figure 1

19 pages, 9131 KiB

Open AccessArticle

Multilayer Semantic Features Adaptive Distillation for Object Detectors

by Zhenchang Zhang, Jinqiang Liu, Yuping Chen, Wang Mei, Fuzhong Huang and Lei Chen

Sensors 2023, 23(17), 7613; https://doi.org/10.3390/s23177613 - 02 Sep 2023

Viewed by 765

Abstract

Knowledge distillation (KD) is a well-established technique for compressing neural networks and has gained increasing attention in object detection tasks. However, typical object detection distillation methods use fixed-level semantic features for distillation, which might not be best for all training stages and samples. In this paper, a multilayer semantic feature adaptive distillation (MSFAD) method is proposed that uses a routing network composed of a teacher and a student detector, along with an agent network for decision making. Specifically, the inputs to the proxy network consist of the features output by the neck structures of the teacher and student detectors, and the output is a decision on which features to choose for distillation. The MSFAD method improves the distillation training process by enabling the student detector to automatically select valuable semantic-level features from the teacher detector. Experimental results demonstrated that the proposed method increased the mAP₅₀ of YOLOv5s by 3.4% and the mAP_50–90 by 3.3%. Additionally, YOLOv5n with only 1.9 M parameters achieved detection performance comparable to that of YOLOv5s. Full article

(This article belongs to the Special Issue Emerging Machine Learning Techniques in Industrial Internet of Things)

► Show Figures

Figure 1

22 pages, 4072 KiB

Open AccessArticle

A Low-Carbon and Economic Dispatch Strategy for a Multi-Microgrid Based on a Meteorological Classification to Handle the Uncertainty of Wind Power

by Yang Liu, Xueling Li and Yamei Liu

Sensors 2023, 23(11), 5350; https://doi.org/10.3390/s23115350 - 05 Jun 2023

Cited by 1 | Viewed by 962

Abstract

In a modern power system, reducing carbon emissions has become a significant goal in mitigating the impact of global warming. Therefore, renewable energy sources, particularly wind-power generation, have been extensively implemented in the system. Despite the advantages of wind power, its uncertainty and randomness lead to critical security, stability, and economic issues in the power system. Recently, multi-microgrid systems (MMGSs) have been considered as a suitable wind-power deployment candidate. Although wind power can be efficiently utilized by MMGSs, uncertainty and randomness still have a significant impact on the dispatching and operation of the system. Therefore, to address the wind power uncertainty issue and achieve an optimal dispatching strategy for MMGSs, this paper presents an adjustable robust optimization (ARO) model based on meteorological clustering. Firstly, the maximum relevance minimum redundancy (MRMR) method and the CURE clustering algorithm are employed for meteorological classification in order to better identify wind patterns. Secondly, a conditional generative adversarial network (CGAN) is adopted to enrich the wind-power datasets with different meteorological patterns, resulting in the construction of ambiguity sets. Thirdly, the uncertainty sets that are finally employed by the ARO framework to establish a two-stage cooperative dispatching model for MMGS can be derived from the ambiguity sets. Additionally, stepped carbon trading is introduced to control the carbon emissions of MMGSs. Finally, the alternative direction method of multipliers (ADMM) and the column and constraint generation (C&CG) algorithm are adopted to achieve a decentralized solution for the dispatching model of MMGSs. Case studies indicate that the presented model has a great performance in improving the wind-power description accuracy, increasing cost efficiency, and reducing system carbon emissions. However, the case studies also report that the approach consumes a relative long running time. Therefore, in future research, the solution algorithm will be further improved for the purpose of raising the efficiency of the solution. Full article

(This article belongs to the Special Issue Emerging Machine Learning Techniques in Industrial Internet of Things)

► Show Figures

Figure 1

17 pages, 1719 KiB

Open AccessArticle

Graph-Based Self-Training for Semi-Supervised Deep Similarity Learning

by Yifan Wang, Yan Huang, Qicong Wang, Chong Zhao, Zhenchang Zhang and Jian Chen

Sensors 2023, 23(8), 3944; https://doi.org/10.3390/s23083944 - 13 Apr 2023

Cited by 1 | Viewed by 1598

Abstract

Semi-supervised learning is a learning pattern that can utilize labeled data and unlabeled data to train deep neural networks. In semi-supervised learning methods, self-training-based methods do not depend on a data augmentation strategy and have better generalization ability. However, their performance is limited by the accuracy of predicted pseudo-labels. In this paper, we propose to reduce the noise in the pseudo-labels from two aspects: the accuracy of predictions and the confidence of the predictions. For the first aspect, we propose a similarity graph structure learning (SGSL) model that considers the correlation between unlabeled and labeled samples, which facilitates the learning of more discriminative features and, thus, obtains more accurate predictions. For the second aspect, we propose an uncertainty-based graph convolutional network (UGCN), which can aggregate similar features based on the learned graph structure in the training phase, making the features more discriminative. It can also output the uncertainty of predictions in the pseudo-label generation phase, generating pseudo-labels only for unlabeled samples with low uncertainty; thus, reducing the noise in the pseudo-labels. Further, a positive and negative self-training framework is proposed, which combines the proposed SGSL model and UGCN into the self-training framework for end-to-end training. In addition, in order to introduce more supervised signals in the self-training process, negative pseudo-labels are generated for unlabeled samples with low prediction confidence, and then the positive and negative pseudo-labeled samples are trained together with a small number of labeled samples to improve the performance of semi-supervised learning. The code is available upon request. Full article

(This article belongs to the Special Issue Emerging Machine Learning Techniques in Industrial Internet of Things)

► Show Figures

Figure 1

22 pages, 1949 KiB

Open AccessArticle

Effects of Particle Swarm Optimisation on a Hybrid Load Balancing Approach for Resource Optimisation in Internet of Things

by Dorcas Dachollom Datiri and Maozhen Li

Sensors 2023, 23(4), 2329; https://doi.org/10.3390/s23042329 - 20 Feb 2023

Cited by 1 | Viewed by 1318

Abstract

The internet of things, a collection of diversified distributed nodes, implies a varying choice of activities ranging from sleep monitoring and tracking of activities, to more complex activities such as data analytics and management. With an increase in scale comes even greater complexities, leading to significant challenges such as excess energy dissipation, which can lead to a decrease in IoT devices’ lifespan. Internet of things’ (IoT) multiple variable activities and ample data management greatly influence devices’ lifespan, making resource optimisation a necessity. Existing methods with respect to aspects of resource management and optimisation are limited in their concern of devices energy dissipation. This paper therefore proposes a decentralised approach, which contains an amalgamation of efficient clustering techniques, edge computing paradigms, and a hybrid algorithm, targeted at curbing resource optimisation problems and life span issues associated with IoT devices. The decentralised topology aimed at the resource optimisation of IoT places equal importance on resource allocation and resource scheduling, as opposed to existing methods, by incorporating aspects of the static (round robin), dynamic (resource-based), and clustering (particle swarm optimisation) algorithms, to provide a solid foundation for an optimised and secure IoT. The simulation constructs five test-case scenarios and uses performance indicators to evaluate the effects the proposed model has on resource optimisation in IoT. The simulation results indicate the superiority of the PSOR2B to the ant colony, the current centralised optimisation approach, LEACH, and C-LBCA. Full article

(This article belongs to the Special Issue Emerging Machine Learning Techniques in Industrial Internet of Things)

► Show Figures

Journal Menu

Journal Browser

Emerging Machine Learning Techniques in Industrial Internet of Things

Share This Special Issue

Special Issue Editors

Special Issue Information

Keywords

Published Papers (7 papers)

Research

Further Information

Guidelines

MDPI Initiatives

Follow MDPI