Training Costs Research Articles

Comparative performance analysis of simple U-Net, residual attention U-Net, and VGG16-U-Net for inventory inland water bodies

Inland water bodies play a vital role at all scales in the terrestrial water balance and Earth’s climate variability. Thus, an inventory of inland waters is crucially important for hydrologic and ecological studies and management. Therefore, the main aim of this study was to develop a deep learning-based method for inventorying and mapping inland water bodies using the RGB band of high-resolution satellite imagery automatically and accurately.The Sentinel-2 Harmonized dataset, together with ZABAGED-validated ground truth, was used as the main dataset for the model training step. Three different deep learning algorithms based on U-Net architecture were employed to segment inland waters, including a simple U-Net, Residual Attention U-Net, and VGG16-U-Net. All three algorithms were trained using a combination of Sentinel-2 visible bands (Red [B04; 665nm], Green [B03; 560nm], and Blue [B02; 490 nm]) at a 10-meter spatial resolution.The Residual Attention U-Net achieved the highest computational cost due to the increased number of trainable parameters. The VGG16-U-Net had the shortest run time and the lowest number of trainable parameters, attributed to its architecture compared to the simple and Residual Attention U-Net architectures, respectively. As a result, the VGG16-U-Net provided the best segmentation results with a mean-IoU score of 0.9850, a slight improvement compared to other proposed U-Net-based architectures.Although the accuracy of the model based on VGG16-U-Net does not make a difference from Residual Attention U-Net, the computation costs for training VGG16-U-Net were dramatically lower than Residual Attention U-Net.

Digital-Twin-Based Fire Safety Management Framework for Smart Buildings

In recent years, the implementation of digital twin (DT) technology has gained significant attention in various industries. However, the fire safety management (FSM) sector has been relatively slow in adopting this technology compared to other major industries. Therefore, this study aims to explore the limitations, opportunities, and challenges associated with adopting DT technology in the FSM sector and further develop a DT-based FSM framework towards smart facility management (FM). To achieve this objective, this research started by reviewing several promising DTs for FSM, including building information modeling (BIM), the Internet of Things (IoT), artificial intelligence (AI), and augmented reality (AR). On this basis, a conceptual framework was synthesized in consideration of the benefits of each technology. A questionnaire was conducted for FM professionals to evaluate the proposed framework and identify the challenges of adopting DT in the FSM sector. The survey results reveal that the proposed framework can assist decision makers in obtaining comprehensive information about facilities’ communication among stakeholders. The survey results validate the potential of the adoption of DTs toward smart FM practices in FSM. The survey results provide insights into the perception of DT technology among FM practitioners and identify the current state of DT technology in the FSM sector, its expected benefits, and its potential challenges. The main barriers to adopting DTs in FSM are a lack of knowledge about DTs, their initial costs, user acceptance, difficulties in systems integration, education training costs, a lack of competence, development complexity, difficulties in data management, and a lack of trust in data security.

A deep learning model for multi-modal spatio-temporal irradiance forecast

Short-term spatio-temporal solar irradiance forecasting plays a pivotal role in scheduling and dispatching energy for distributed energy systems. Fluctuations in cloud cover can be monitored via satellite cloud imagery, which directly impacts irradiance. However, integrating and fusing multi-source heterogeneous data, such as satellite cloud images and ground monitoring data from distributed stations, remains challenging. Here, a spatio-temporal irradiance forecast model is proposed based on multi-modal deep learning model to predict global horizontal irradiance 30 min ahead. To address the feature extraction of heterogeneous data, a dual-channel structure consisting of a time-series processing block and a satellite cloud image processing block is developed to enable parallel processing of multi-modal features. In order To tightly couple cloud images and historical time series at the feature level, maximum mean discrepancy of these two feature is used to help the fusion of heterogeneous data. Furthermore, a self-attention mechanism is employed to construct adaptive inter-region information weights to enhance spatio-temporal representation ability. The evaluation of the method is conducted on open-access datasets from six locations in Jiangsu Province, China. Experimental results demonstrate that the proposed model efficiently utilizes heterogeneous data to improve prediction accuracy under various conditions and enhances model robustness, reducing RMSE by 2.8%–20.58%. Meanwhile, the proposed end-to-end model reduces training and deployment costs for real-world use.

Electrochemical Impedance Spectrum Equivalent Circuit Parameter Identification Using a Deep Learning Technique

Physical models are suitable for the development and optimization of materials and cell designs, whereas models based on experimental data and electrical equivalent circuits (EECs) are suitable for the development of operation estimators, both for cells and batteries. This research work develops an innovative unsupervised artificial neural network (ANN) training cost function for identifying equivalent circuit parameters using electrochemical impedance spectroscopy (EIS) to identify and monitor parameter variations associated with different physicochemical processes that can be related to the states or failure modes in batteries. Many techniques and algorithms are used to fit a predefined EEC parameter, many requiring high-human-expertise support work. However, once the appropriate EEC model is selected to model the different physicochemical processes associated with a given battery technology, the challenge is to implement algorithms that can automatically calculate parameter variations in real time to allow the implementation of estimators of capacity, health, safety, and other degradation modes. Based on previous studies using data augmentation techniques, the new ANN deep learning method introduced in this study yields better results than classical training algorithms. The data used in this work are based on an aging and characterization dataset for 80 Ah and 12 V lead–acid batteries.

Robots and Artificial Intelligence in the Military.

The Fourth Industrial Revolution introduced numerous disruptive technologies that are changing human life faster than ever before. Right on its heel, the Fifth Industrial Revolution is introducing robotics using Artificial Intelligence aided by Machine Learning. While these are considered the future, the disruption caused by them is forcing changes in how we operate and perceive things. One such area is the ‘military’. Though Artificial Intelligence and robotics are being experimented with as a replacement for humans to de-risk their lives, they are found to be advantageous to reduce the cost of training, rations, supply and maintenance of personnel. They are thus receiving increasing investments for development and deployment. This in return is altering how warfighting is perceived and conducted. To ensure a smooth progression of humans with the numerous technological developments being made in the field of Artificial Intelligence, the article looks at the advances made in these fields for military applications that are improving their sustainability and control, the world over. In addition, to assist decision-makers, a broad-based roadmap for the use of Artificial Intelligence in the military is discussed.

Availability of basic first aid training for cardiac arrest for people with disabilities

Aim. Basic first aid digital training for cardiac arrest opens up opportunities for mass dissemination of knowledge, popularization and motivation of the population to provide assistance, and can also replace the theoretical part of classroom training, reducing the cost of training. However, due to poor design, such training may not be available to people with disabilities (PWD), who constitute a large population of more than one billion potential participants in first aid. The aim was to study the digital accessibility (web accessibility) of massive open online courses in cardiopulmonary resuscitation (CPR) for PWD.Material and methods. A search was carried out for free English-language online courses teaching the principles and rules of providing first aid in cardiac arrest. In addition, the digital content of online courses was assessed for compliance with the international Web Content Accessibility Guidelines (WCAG) 2.1.Results. All basic CPR online courses analyzed (n=28) were characterized by poor web accessibility. For each course, between 26,9% and 51,3% of the WCAG 2.1 guideline success criteria were not met. All courses have web accessibility errors related to text contrast and visual issues, lack of alternative text descriptions for non-text content, missing abbreviations, and web page layout issues that prevent the use of assistive technologies (such as text-to-speech software). Overall, none of the courses meet all of the criteria for the minimum acceptable level of compliance with WCAG 2.1 guidelines. The discovered problems of web accessibility can serve as serious, sometimes insurmountable barriers to the mastery of courses by students (including older people) with visual impairments, hearing loss, motor, cognitive and neurological disorders.Conclusion. All free mass online courses on basic CPR have deficiencies in the availability of digital content, which significantly complicate the learning of educational materials by PWD. The development of uniform guidelines for accessible educational content, along with the intensification of research and the implementation of organizational and methodological measures to ensure the availability of educational resources on first aid for various categories of the population, should help create equal opportunities for quality training. In addition, this should significantly expand the population coverage with training, and increase the frequency and effectiveness of first aid.

FACTORS INFLUENCING EMPLOYEE RETENTION AMONG LECTURERS IN HIGHER EDUCATION INSTITUTIONS

Department of Business Management and Accounting, Kolej Poly-Tech MARA Alor Setar, Malaysia Employee turnover among lecturers in Malaysia is a significant issue that can result in a loss of talent and experience, decreased productivity, and increased costs for recruitment and training of new staff. The literature suggests that low salaries and benefits and a lack of career growth opportunities are the main reasons for high employee turnover among lecturers in Malaysia. These factors can lead to a lack of job satisfaction, and motivation, and a feeling of career stagnation, resulting in many lecturers intending to leave their current institutions. This paper aims to present a conceptual paper on the factors influencing employee retention among lecturers in Malaysia. The study draws on various sources, including academic studies and surveys, to identify the main factors contributing to employee turnover and the costs associated with it. This paper also provides insights into the challenges faced by universities in retaining their lecturers and the need to implement strategies to improve retention. The study highlights the importance of addressing the issues of low salaries and benefits and the lack of career growth opportunities to improve lecturer retention in Malaysia. The study suggests that universities need to develop retention strategies that offer competitive salaries, benefits, and opportunities for career growth to retain their lecturers. By doing so, universities can reduce the costs associated with employee turnover and ensure that they have experienced and talented lecturers to deliver high-quality education and research.

ADGNN: Towards Scalable GNN Training with Aggregation-Difference Aware Sampling

Distributed computing is promising to enable large-scale graph neural network (GNN) model training. However, care is needed to avoid excessive computational and communication overheads. Sampling is promising in terms of enabling scalability, and sampling techniques have been proposed to reduce training costs. However, online sampling introduces large overheads, and while offline sampling that is done only once can eliminate such overheads, it instead introduces information loss and accuracy degradation. Thus, existing sampling techniques are unable to improve simultaneously both efficiency and accuracy, particularly at low sampling rates. We develop a distributed system, ADGNN, for full-batch based GNN training that adopts a hybrid sampling architecture to enable a trade-off between efficiency and accuracy. Specifically, ADGNN employs sampling result reuse techniques to reduce the cost associated with sampling and thus improve training efficiency. To alleviate accuracy degradation, we introduce a new metric,Aggregation Difference (AD), that quantifies the gap between sampled and full neighbor set aggregation. We present so-called AD-Sampling that aims to minimize the Aggregation Difference with an adaptive sampling frequency tuner. Finally, ADGNN employs anAD -importance-based sampling technique for remote neighbors to further reduce communication costs. Experiments on five real datasets show that ADGNN is able to outperform the state-of-the-art by up to nearly 9 times in terms of efficiency, while achieving comparable accuracy to the non-sampling methods.

ОСОБЕННОСТИ ОБУЧЕНИЯ ПЕРСОНАЛА В 2023 ГОДУ

The article discusses the meaning and essence of personnel training, its main types. Based on the research results, the features of the organization of training in 2023 were identified. The priority directions of companies’ development and their strategies are analyzed. The article discusses training both from the employer and employees, which is reflected in the opinion of employees about possible changes after training. The economic part is also important, so the article considers the opinion of respondents (companies) about the change in the budget for training in 2023. The article also discusses two examples of employee development programs in large companies in order to reduce the cost of staff training, in formats such as distance learning and corporate university.

TrueDeep: A systematic approach of crack detection with less data

Supervised and semi-supervised semantic segmentation algorithms require significant amount of annotated data to achieve a good performance. In many situations, the data is either not available or the annotation is expensive. The objective of this work is to show that by incorporating domain knowledge along with deep learning architectures, we can achieve similar performance with less data. We have used publicly available crack segmentation datasets and shown that selecting the input images using knowledge can significantly boost the performance of deep-learning based architectures. Our proposed approaches have many fold advantages such as low annotation and training cost, and less energy consumption. We have measured the performance of our algorithm quantitatively in terms of mean intersection over union (mIoU) and F-score. Our representative model, developed with 23% of the overall data; surpasses the baseline model (>1.5% F-score gain) on the test data while matching the benchmark performance and, with a suitable augmentation strategy, outperforms the baseline model (11.5% average F-score gain) and the benchmark (>8.75% average F-score gain) over the selected blind datasets.

Cost of simulation-based mastery learning for abdominal ultrasound

BackgroundUltrasound is an essential diagnostic examination used in several medical specialties. However, the quality of ultrasound examinations is dependent on mastery of certain skills, which may be difficult and costly to attain in the clinical setting. This study aimed to explore mastery learning for trainees practicing general abdominal ultrasound using a virtual reality simulator and to evaluate the associated cost per student achieving the mastery learning level.MethodsTrainees were instructed to train on a virtual reality ultrasound simulator until the attainment of a mastery learning level was established in a previous study. Automated simulator scores were used to track performances during each round of training, and these scores were recorded to determine learning curves. Finally, the costs of the training were evaluated using a micro-costing procedure.ResultsTwenty-one out of the 24 trainees managed to attain the predefined mastery level two times consecutively. The trainees completed their training with a median of 2h38min (range: 1h20min-4h30min) using a median of 7 attempts (range: 3–11 attempts) at the simulator test. The cost of training one trainee to the mastery level was estimated to be USD 638.ConclusionComplete trainees can obtain mastery learning levels in general abdominal ultrasound examinations within 3 hours of training in the simulated setting and at an average cost of USD 638 per trainee. Future studies are needed to explore how the cost of simulation-based training is best balanced against the costs of clinical training.

Research on water extraction from high resolution remote sensing images based on deep learning

Introduction: Monitoring surface water through the extraction of water bodies from high-resolution remote sensing images is of significant importance. With the advancements in deep learning, deep neural networks have been increasingly applied to high-resolution remote sensing image segmentation. However, conventional convolutional models face challenges in water body extraction, including issues like unclear water boundaries and a high number of training parameters.Methods: In this study, we employed the DeeplabV3+ network for water body extraction in high-resolution remote sensing images. However, the traditional DeeplabV3+ network exhibited limitations in segmentation accuracy for high-resolution remote sensing images and incurred high training costs due to a large number of parameters. To address these issues, we made several improvements to the traditional DeeplabV3+ network: Replaced the backbone network with MobileNetV2. Added a Channel Attention (CA) module to the MobileNetV2 feature extraction network. Introduced an Atrous Spatial Pyramid Pooling (ASPP) module. Implemented Focal loss for balanced loss computation.Results: Our proposed method yielded significant enhancements. It not only improved the segmentation accuracy of water bodies in high-resolution remote sensing images but also effectively reduced the number of network parameters and training time. Experimental results on the Water dataset demonstrated superior performance compared to other networks: Outperformed the U-Net network by 3.06% in terms of mean Intersection over Union (mIoU). Outperformed the MACU-Net network by 1.03%. Outperformed the traditional DeeplabV3+ network by 2.05%. The proposed method surpassed not only the traditional DeeplabV3+ but also U-Net, PSP-Net, and MACU-Net networks.Discussion: These results highlight the effectiveness of our modified DeeplabV3+ network with MobileNetV2 backbone, CA module, ASPP module, and Focal loss for water body extraction in high-resolution remote sensing images. The reduction in training time and parameters makes our approach a promising solution for accurate and efficient water body segmentation in remote sensing applications.

Generating intelligent agent behaviors in multi-agent game AI using deep reinforcement learning algorithm

<span>The utilization of games in training the reinforcement learning (RL) agent is to describe the complex and high-dimensional real-world data. By utilizing games, RL researchers will be able to evade high experimental costs in training an agent to do intelligence tasks. The objective of this research is to generate intelligent agent behaviors in multi-agent game artificial intelligence (AI) using deep reinforcement learning (DRL) algorithm. A basic RL algorithm called deep Q network is chosen to be implemented. The agent is trained by the environment's raw pixel images and the action list information. The experiments conducted by using this algorithm show the agent’s decision-making ability in choosing a favorable action. In the default setting for the algorithm, the training is set into 1 epoch and 0.0025 learning rate. The number of training iterations is set to one as the training function will be repeatedly called for every 4-timestep. However, the author also experimented with two different scenarios in training the agent and compared the results. The experimental findings demonstrate that our agents learn correctly and successfully while actively participating in the game in real time. Additionally, our agent can quickly adjust against a different enemy on a varied map because of the observed knowledge from prior training.</span>

Domain independent post-processing with graph U-nets: applications to electrical impedance tomographic imaging⋆ ⋆SH and BH were supported by the National Institute Of Biomedical Imaging And Bioengineering of the National Institutes of Health under Award Number R21EB028064. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health. AH

Objective. To extend the highly successful U-Net Convolutional Neural Network architecture, which is limited to rectangular pixel/voxel domains, to a graph-based equivalent that works flexibly on irregular meshes; and demonstrate the effectiveness on electrical impedance tomography (EIT). Approach. By interpreting the irregular mesh as a graph, we develop a graph U-Net with new cluster pooling and unpooling layers that mimic the classic neighborhood based max-pooling important for imaging applications. Main results. The proposed graph U-Net is shown to be flexible and effective for improving early iterate total variation (TV) reconstructions from EIT measurements, using as little as the first iteration. The performance is evaluated for simulated data, and on experimental data from three measurement devices with different measurement geometries and instrumentations. We successfully show that such networks can be trained with a simple two-dimensional simulated training set, and generalize to very different domains, including measurements from a three-dimensional device and subsequent 3D reconstructions. Significance. As many inverse problems are solved on irregular (e.g. finite element) meshes, the proposed graph U-Net and pooling layers provide the added flexibility to process directly on the computational mesh. Post-processing an early iterate reconstruction greatly reduces the computational cost which can become prohibitive in higher dimensions with dense meshes. As the graph structure is independent of ‘dimension’, the flexibility to extend networks trained on 2D domains to 3D domains offers a possibility to further reduce computational cost in training.

Vertical Layering of Quantized Neural Networks for Heterogeneous Inference.

Although considerable progress has been obtained in neural network quantization for efficient inference, existing methods are not scalable to heterogeneous devices as one dedicated model needs to be trained, transmitted, and stored for one specific hardware setting, incurring considerable costs in model training and maintenance. In this paper, we study a new vertical-layered representation of neural network weights for encapsulating all quantized models into a single one. It represents weights as a group of bits (vertical layers) organized from the most significant bit (also called the basic layer) to less significant bits (enhance layers). Hence, a neural network with an arbitrary quantization precision can be obtained by adding corresponding enhance layers to the basic layer. However, we empirically find that models obtained with existing quantization methods suffer severe performance degradation if adapted to vertical-layered weight representation. To this end, we propose a simple once quantization-aware training (QAT) scheme for obtaining high-performance vertical-layered models. Our design incorporates a cascade downsampling mechanism with the multi-objective optimization employed to train the shared source model weights such that they can be updated simultaneously, considering the performance of all networks. After the model is trained, to construct a vertical-layered network, the lowest bit-width quantized weights become the basic layer, and every bit dropped along the downsampling process act as an enhance layer. Our design is extensively evaluated on CIFAR-100 and ImageNet datasets. Experiments show that the proposed vertical-layered representation and developed once QAT scheme are effective in embodying multiple quantized networks into a single one and allow one-time training, and it delivers comparable performance as that of quantized models tailored to any specific bit-width.

A Multi-Perspective Feature Approach to Few-Shot Classification of IoT Traffic

The realization of Internet of Things (IoT) traffic classification is crucial to the management and monitoring of the IoT network. Performing IoT traffic classification under a few-shot scenario is vital due to extremely rapid growth in the number of IoT devices and necessity to save computational costs. In this paper, we propose MAFFIT, a multi-perspective feature approach to few-shot classification of IoT traffic. The purpose is to comprehensively consider the information of network traffic and to achieve accurate classification of IoT traffic using a limited number of samples. MAFFIT is based on our key observation that traffic behaviour and traffic composition are highly consistent across IoT traffic of the same class. For a flow, MAFFIT will first extract the packet length sequences and packet byte sequence, then encodes the features of the corresponding sequences using feature construction, and finally uses comparative learning to obtain the class of the flow without the additional cost of training a comparison model. We conduct extensive experiments on two real-world IoT traffic datasets, the results demonstrate that MAFFIT can achieve accurate IoT traffic classification using a limited number of flow samples and MAFFIT outperforms three existing network traffic classification methods.

GhostFormer: Efficiently amalgamated CNN-transformer architecture for object detection

The lightweight network model has gradually evolved into an important research direction in object detection. Network lightweight design has a variety of research methods, such as quantization, knowledge distillation, and neural architecture search. However, these methods either fail to break through the performance bottleneck of the model itself or require massive training costs. In order to solve these problems, a new object detection model based on CNN-Transformer hybrid feature extraction network called GhostFormer is proposed from the perspective of lightweight network structure design. GhostFormer makes full use of the advantages of local modeling of CNN and global modeling of Transformer, not only effectively reducing the complexity of the convolution model but also breaking through the limitation of Transformer’s lack of inductive bias. Finally, better transfer results are obtained in downstream tasks. Experiments show that the model is less than half as computationally expensive as YOLOv7 on the Pascal VOC dataset, with only about 3 % mAP@0.5 loss, and 9.7% mAP@0.5:0.95 improvement on the MS COCO dataset compared with GhostNet.

Use of Secondary Rail Lines as Alternative Routes for Freight Rail Transportation

Abstract Rail transport is currently facing a number of problems. The European Union wants to shift as much freight as possible from road to rail and this is one of the key objectives of the Green Deal. This objective also requires rail carriers to improve their service quality in order to retain existing customers and attract new customers currently using road transport. However, there are a number of factors that are making the railways less competitive. One of these is the closure of the main lines. These lines are crucial to the successful organisation of transport. The article looks at the train costs incurred by carriers when main lines cannot be used for various reasons (line reconstruction, lack of capacity, etc.). The article compares the cost of a train running on a main line with two alternative lines. It examines and describes in detail the efficiency and economic impact of using the alternative routes in the event of an emergency on the main line, and also identifies the advantages and disadvantages of the secondary routes. The paper provides an insight into the usability of secondary routes when an alternative route is required.

Embankment surface crack pixel-wise identification in UAV images based on a lightweight U-Network with transfer learning

Embankments in challenging environments experience cyclic external influences and loads that render their surfaces susceptible to a range of defects, with cracks particularly menacing. Therefore, accurate and efficient crack identification holds paramount significance for embankment safety. This paper addresses the challenges of low crack recognition rates on complex embankment surfaces and the high training costs associated with deep learning-based crack detection methods. We introduce a novel residual structure as the foundational innovation, followed by developing a streamlined U-shaped network (denoted as U²-Net_Aggregation) that amalgamates multi-scale insights by integrating skip connections. The newly devised U²-Net_Aggregation network demonstrates significant improvements in accuracy, intersection over union (IOU), and F1-score metrics through ablation tests on three distinct crack datasets. Comparisons with commonly employed crack detection methods (FCN, SegNet, U-Net, DeepCrack, and U²-Net) were conducted at the pixel level using an embankment crack dataset. Results unequivocally showcase enhancements in both accuracy and processing speed. To overcome the challenge of a high application threshold resulting from a scarcity of training data specific to embankment cracks, we compile an extensive dataset featuring diverse building crack scenarios, serving as a foundation for transfer learning. Following transfer learning, U2-Net_Aggregation attains recall, IOU, and F1-score values of 87.31%, 81.76%, and 90.06%, respectively. The proposed methodology enables swift and accurate crack detection by directly utilizing Unmanned Aerial Vehicle (UAV) images, offering a novel avenue for embankment crack detection in real-world field scenarios.

A Two-Stage Evolutionary Game Model for Collaborative Emergency Management Between Local Governments and Enterprises

Enterprises play a vital role in emergency management, but few studies have considered the strategy choices behind such participation or the collaborative relationship with the government. This study contended that enterprises have at least three strategies regarding emergency management: non-participation, short-term participation, and long-term participation. We constructed a two-stage evolutionary game model to explore the behavioral evolution rules and evolutionary stability strategies of the government and enterprises, and employed numerical simulation to analyze how various factors influence the strategy selection of the government and enterprises. The results show that if and only if the utility value of participation is greater than 0, an enterprise will participate in emergency management. The evolutionary game then enters the second stage, during which system stability is affected by a synergistic relationship between participation cost, reputation benefit, and government subsidies, and by an incremental relationship between emergency management benefit, government subsidies, and emergency training cost. This study provides a new theoretical perspective for research on collaborative emergency management, and the results provide important references for promoting the performance of collaborative emergency management.