Article
Water Resources
M. T. Vu, A. Jardani
Summary: This paper proposes a new approach for simultaneously mapping the fracture/conduit network and the equivalent transmissivity of the rock matrix using hydraulic head measurements. A multitask neural network is used to approximate the joint inversion operator that links hydraulic head data to aquifer hydraulic properties. The network consists of two single-task neural networks, one for fracture structure inversion and the other for transmissivity inversion, and is trained with a large database of synthetic aquifer models.
ADVANCES IN WATER RESOURCES
(2023)
Article
Oncology
Jingjing Yu, Chenyang Dai, Xuelei He, Hongbo Guo, Siyu Sun, Ying Liu
Summary: The paper introduces a deep-learning optical reconstruction method based on 1DCNN to improve the accuracy and efficiency of BLT reconstruction. By establishing the nonlinear mapping relationship between photon flux density and bioluminescence source distribution, it avoids solving the ill-posed inverse problem iteratively, reducing training parameters and improving model learning efficiency. Simulation and in vivo experimental results demonstrate the superiority and stability of the 1DCNN method in practical applications.
FRONTIERS IN ONCOLOGY
(2021)
Article
Computer Science, Artificial Intelligence
Mohammad Hassan Tayarani Najaran
Summary: In this paper, various transform functions, including Fourier and Wavelet, were applied to extract features from vibration data for fault diagnosis. Different learning algorithms were trained for each feature extraction approach and their results were aggregated in an ensemble machine learning algorithm. An evolutionary algorithm was introduced to optimize the weights of the base learner algorithms and find the best architecture for Convolutional Neural Networks (CNNs) in fault diagnosis. Experimental results on benchmark problems were presented to evaluate the proposed algorithm.
EXPERT SYSTEMS WITH APPLICATIONS
(2023)
Article
Automation & Control Systems
Wenqiang Li, Yuk Ming Tang, Ziyang Wang, Kai Ming Yu, Suet To
Summary: Automatic vertebrae segmentation using CT plays a crucial role in automated spine analysis, and recent advancements in deep learning have led to precise performance through deep convolutional neural networks. While DCNN-based semantic segmentation algorithms have advantages, they face limitations that are addressed by the proposed novel algorithm, which includes encoder-decoder framework, Layer Normalization, Atrous Residual Path, and a 3D Attention Module to improve segmentation accuracy. Experimental results show competitive performance compared to existing methods for automatic vertebrae semantic segmentation.
ENGINEERING APPLICATIONS OF ARTIFICIAL INTELLIGENCE
(2022)
Article
Physics, Multidisciplinary
Thilo Strauss, Taufiquar Khan
Summary: Electrical impedance tomography (EIT) is a non-invasive imaging modality used for estimating the conductivity of an object W from boundary electrode measurements. In this paper, we propose a novel and simple neural network architecture for solving the EIT inverse problem using machine learning. The proposed model addresses previous difficulties, including instability, and improves gradient-based methods for estimating the unknown conductivity.
Article
Computer Science, Artificial Intelligence
Aksh Garg, Sana Salehi, Marianna La Rocca, Rachael Garner, Dominique Duncan
Summary: This paper utilizes 20 convolutional neural networks to classify patients as COVID-19 positive, healthy, or suffering from other pulmonary infections based on chest CT scans. The study finds that the EfficientNet-B5 model performs the best, offering a rapid and accurate diagnostic for COVID-19.
EXPERT SYSTEMS WITH APPLICATIONS
(2022)
Article
Engineering, Electrical & Electronic
Runren Zhang, Qingtao Sun, Yiqian Mao, Liangze Cui, Yongze Jia, Wei-Feng Huang, Mohsen Ahmadian, Qing Huo Liu
Summary: This article discusses the application of deep transfer learning in hydraulic fracture imaging. By using a two-step approach, training a convolutional neural network with approximated field patterns generated from a simplified model and fine-tuning it with true field patterns from a full model, accurate reconstruction results can be achieved.
IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION
(2022)
Article
Computer Science, Artificial Intelligence
Tianyu Ma, Alan Q. Wang, Adrian V. Dalca, Mert R. Sabuncu
Summary: The convolutional neural network (CNN) is a commonly used architecture for computer vision tasks. A new building block called hyper-convolution is presented in this paper, which encodes the convolutional kernel using spatial coordinates and enables a more flexible architecture design. Experimental results showed that replacing regular convolutions with hyper-convolutions improved performance with fewer parameters and increased robustness against noise.
MEDICAL IMAGE ANALYSIS
(2023)
Article
Engineering, Biomedical
Emil Y. Sidky, Iris Lorente, Jovan G. Brankov, Xiaochuan Pan
Summary: This study investigates the use of CNN for solving the inverse problem of image reconstruction in sparse-view CT, finding that CNN is unable to accurately recover images in this context but constrained total-variation minimization can. This raises doubts on similar unsupported claims regarding the use of CNNs and deep-learning for solving inverse problems in medical imaging.
IEEE TRANSACTIONS ON BIOMEDICAL ENGINEERING
(2021)
Article
Automation & Control Systems
Susanna Lange, Kyle Helfrich, Qiang Ye
Summary: Batch normalization (BN) is a widely used method in deep learning that has shown success in reducing training time and improving generalization performance. However, it lacks theoretical understanding. In this paper, a new method called Batch Normalization Preconditioning (BNP) is proposed, which applies normalization by conditioning the parameter gradients directly during training, improving the convergence of the loss function's Hessian matrix. BNP has the advantage of not being constrained by mini-batch size and is applicable to online learning. Additionally, the connection between BNP and BN provides theoretical insights on BN's improvement of training and its application to special architectures.
JOURNAL OF MACHINE LEARNING RESEARCH
(2022)
Article
Engineering, Multidisciplinary
Xu-Hui Zhou, Jiequn Han, Heng Xiao
Summary: Constitutive and closure models are essential in computational mechanics and physics, describing the relationship between stress and strain in materials. A neural network-based nonlocal constitutive model has been proposed, showing promising predictive capability and interpretability.
COMPUTER METHODS IN APPLIED MECHANICS AND ENGINEERING
(2021)
Article
Computer Science, Artificial Intelligence
Jialin Liu, Fei Chao, Chih-Min Lin, Changle Zhou, Changjing Shang
Summary: This paper introduces dynamic kernel convolutional neural networks (DK-CNNs) and explains how they enhance the expressive capacity of convolutional operations by extending a latent dimension. DK convolution analyzes fixed features with a latent variable, leading to better performance compared to regular CNNs.
Article
Biology
Jose Denes Lima Araujo, Luana Batista da Cruz, Joao Otavio Bandeira Diniz, Jonnison Lima Ferreira, Aristofanes Correa Silva, Anselmo Cardoso de Paiva, Marcelo Gattass
Summary: This study shows that liver segmentation, even in the presence of lesions in CT images, can be efficiently carried out using a cascade approach and incorporating a reconstruction step based on deep convolutional neural networks.
COMPUTERS IN BIOLOGY AND MEDICINE
(2022)
Article
Computer Science, Information Systems
Hanyue Xu, Kah Phooi Seng, Li-Minn Ang
Summary: This paper proposes a hybrid approach that integrates graph convolutional networks (GCNs) and deep convolutional neural networks (DCNNs) for game strategies. Experimental results show that the hybrid model outperforms the traditional DCNN model in extracting game strategies.
Article
Engineering, Biomedical
Yang Lv, Chen Xi
Summary: In this study, a deep progressive learning (DPL) method is proposed for PET image reconstruction to bridge the gap between low quality and high quality images through two learning steps. The experimental results show promising outcomes in reducing noise and improving contrast of PET images, with potential versatility for various imaging and image processing problems.
PHYSICS IN MEDICINE AND BIOLOGY
(2021)
Article
Engineering, Civil
Arfan Arshad, Ali Mirchi, Javier Vilcaez, Muhammad Umar Akbar, Kaveh Madani
Summary: High-resolution, continuous groundwater data is crucial for adaptive aquifer management. This study presents a predictive modeling framework that incorporates covariates and existing observations to estimate groundwater level changes. The framework outperforms other methods and provides reliable estimates for unmonitored sites. The study also examines groundwater level changes in different regions and highlights the importance of effective aquifer management.
JOURNAL OF HYDROLOGY
(2024)
Article
Engineering, Civil
Lihua Chen, Jie Deng, Wenzhe Yang, Hang Chen
Summary: A new grid-based distributed karst hydrological model (GDKHM) is developed to simulate streamflow in the flood-prone karst area of Southwest China. The results show that the GDKHM performs well in predicting floods and capturing the spatial variability of karst system.
JOURNAL OF HYDROLOGY
(2024)
Article
Engineering, Civil
Faruk Gurbuz, Avinash Mudireddy, Ricardo Mantilla, Shaoping Xiao
Summary: Machine learning algorithms have shown better performance in streamflow prediction compared to traditional hydrological models. In this study, researchers proposed a methodology to test and benchmark ML algorithms using artificial data generated by physically-based hydrological models. They found that deep learning algorithms can correctly identify the relationship between streamflow and rainfall in certain conditions, but fail to outperform traditional prediction methods in other scenarios.
JOURNAL OF HYDROLOGY
(2024)
Article
Engineering, Civil
Yadong Ji, Jianyu Fu, Bingjun Liu, Zeqin Huang, Xuejin Tan
Summary: This study distinguishes the uncertainty in drought projection into scenario uncertainty, model uncertainty, and internal variability uncertainty. The results show that the estimation of total uncertainty reaches a minimum in the mid-21st century and that model uncertainty is dominant in tropical regions.
JOURNAL OF HYDROLOGY
(2024)
Article
Engineering, Civil
Z. R. van Leeuwen, M. J. Klaar, M. W. Smith, L. E. Brown
Summary: This study quantifies the effectiveness of leaky dams in reducing flood peak magnitude using a transfer function noise modelling approach. The results show that leaky dams have a significant but highly variable impact on flood peak magnitude, and managing expectations should consider event size and type.
JOURNAL OF HYDROLOGY
(2024)
Article
Engineering, Civil
Zeda Yin, Yasaman Saadati, M. Hadi Amini, Linlong Bian, Beichao Hu
Summary: Combined sewer overflows pose significant threats to public health and the environment, and various strategies have been proposed to mitigate their adverse effects. Smart control strategies have gained traction due to their cost-effectiveness but face challenges in balancing precision and computational efficiency. To address this, we propose exploring machine learning models and the inversion of neural networks for more efficient CSO prediction and optimization.
JOURNAL OF HYDROLOGY
(2024)
Article
Engineering, Civil
Qimou Zhang, Jiacong Huang, Jing Zhang, Rui Qian, Zhen Cui, Junfeng Gao
Summary: This study developed a N-cycling model for lowland rural rivers covered by macrophytes and investigated the N imports, exports, and response to sediment dredging. The findings showed a considerable N retention ability in the study river, with significant N imports from connected rivers and surrounding polders. Sediment dredging increased particulate nitrogen resuspension and settling rates, while decreasing ammonia nitrogen release, denitrification, and macrophyte uptake rates.
JOURNAL OF HYDROLOGY
(2024)
Article
Engineering, Civil
Xue Li, Yingyin Zhou, Jian Sha, Man Zhang, Zhong-Liang Wang
Summary: High-resolution climate data is crucial for predicting regional climate and water environment changes. In this study, a two-step downscaling method was developed to enhance the spatial resolution of GCM data and improve the accuracy for small basins. The method combined medium-resolution climate data with high-resolution topographic data to capture spatial and temporal details. The downscaled climate data were then used to simulate the impacts of climate change on hydrology and water quality in a small basin. The results demonstrated the effectiveness of the downscaling method for spatially differentiated simulations.
JOURNAL OF HYDROLOGY
(2024)
Article
Engineering, Civil
Tongqing Shen, Peng Jiang, Jiahui Zhao, Xuegao Chen, Hui Lin, Bin Yang, Changhai Tan, Ying Zhang, Xinting Fu, Zhongbo Yu
Summary: This study evaluates the long-term interannual dynamics of permafrost distribution and active layer thickness on the Tibetan Plateau, and predicts future degradation trends. The results show that permafrost area has been decreasing and active layer thickness has been increasing, with an accelerated degradation observed in recent decades. This has significant implications for local water cycle processes, water ecology, and water security.
JOURNAL OF HYDROLOGY
(2024)
Article
Engineering, Civil
Chi Zhang, Xu Zhang, Qiuhong Tang, Deliang Chen, Jinchuan Huang, Shaohong Wu, Yubo Liu
Summary: Precipitation over the Tibetan Plateau is influenced by systems such as the Asian monsoons, the westerlies, and local circulations. The Indian monsoon, the westerlies, and local circulations are the main systems affecting precipitation over the entire Tibetan Plateau. The East Asian summer monsoon primarily affects the eastern Tibetan Plateau. The Indian monsoon has the greatest influence on precipitation in the southern and central grid cells, while the westerlies have the greatest influence on precipitation in the northern and western grid cells. Local circulations have the strongest influence on the central and eastern grid cells.
JOURNAL OF HYDROLOGY
(2024)
Article
Engineering, Civil
Manuel Almeida, Antonio Rodrigues, Pedro Coelho
Summary: This study aimed to improve the accuracy of Total Phosphorus export coefficient models, which are essential for water management. Four different models were applied to 27 agroforestry watersheds in the Mediterranean region. The modeling approach showed significant improvements in predicting the Total Phosphorus diffuse loads.
JOURNAL OF HYDROLOGY
(2024)
Article
Engineering, Civil
Yutao Wang, Haojie Yin, Ziyi Wang, Yi Li, Pingping Wang, Longfei Wang
Summary: This study investigated the distribution and transformation of dissolved organic nitrogen (DON) in riverbed sediments impacted by effluent discharge. The authors found that the spectral characteristics of dissolved organic matter (DOM) in surface water and sediment porewater could be used to predict DON variations in riverbed sediments. Random forest and extreme gradient boosting machine learning methods were employed to provide accurate predictions of DON content and properties at different depths. These findings have important implications for wastewater discharge management and river health.
JOURNAL OF HYDROLOGY
(2024)
Article
Engineering, Civil
Saba Mirza Alipour, Kolbjorn Engeland, Joao Leal
Summary: This study assesses the uncertainty associated with 100-year flood maps under different scenarios using Monte Carlo simulations. The findings highlight the importance of employing probabilistic approaches for accurate and secure flood maps, with the selection of probability distribution being the primary source of uncertainty in precipitation.
JOURNAL OF HYDROLOGY
(2024)
Article
Engineering, Civil
Janine A. de Wit, Marjolein H. J. van Huijgevoort, Jos C. van Dam, Ge A. P. H. van den Eertwegh, Dion van Deijl, Coen J. Ritsema, Ruud P. Bartholomeus
Summary: The study focuses on the hydrological consequences of controlled drainage with subirrigation (CD-SI) on groundwater level, soil moisture content, and soil water potential. The simulations show that CD-SI can improve hydrological conditions for crop growth, but the success depends on subtle differences in geohydrologic characteristics.
JOURNAL OF HYDROLOGY
(2024)
Article
Engineering, Civil
Constantin Seidl, Sarah Ann Wheeler, Declan Page
Summary: Water availability and quality issues will become increasingly important in the future due to climate change impacts. Managed Aquifer Recharge (MAR) is an effective water management tool, but often overlooked. This study analyzes global MAR applications and identifies the key factors for success, providing valuable insights for future design and application.
JOURNAL OF HYDROLOGY
(2024)