Article
Computer Science, Information Systems
Gwo-Jiun Horng, Min-Xiang Liu, Chien-Chin Hsu
Summary: The treatment of brain ischemia with tissue plasminogen activator may lead to brain hemorrhage risk; decision-making under rescue time pressure is crucial. Lack of benchmarks due to uncertainties post-treatment; Utilization of adaptive deep autoencoder model to learn features of specific data. Preprocessing of data proposed with methods like K-means and image denoising for maximum area preservation; Use of VAE WGAN-GP to generate 3D medical images for insufficient training data, with focus on real data preprocessing and image generation techniques.
COMPUTER COMMUNICATIONS
(2021)
Article
Engineering, Chemical
Yifan Liu, Runze Liu, Jinyu Duan, Li Wang, Xiangwen Zhang, Guozhu Li
Summary: Commercial fuel discovery is facing decreasing return of investment due to stricter environmental criteria and reducing potential uses for each new fuel. This study proposes a deep generative model called LIGANDS, which screens desired fuel molecules in a large chemical space without manually setting design rules. LIGANDS integrates a variational autoencoder, a generative adversarial network, and a stacking model to generate new fuel molecules with similar properties and improved energy performance. The model imitates key properties of target fuel to expand and enrich the fuel-relevant chemical space with innovative molecular entities.
CHEMICAL ENGINEERING SCIENCE
(2023)
Article
Chemistry, Multidisciplinary
Yuxia Tang, Jiulou Zhang, Doudou He, Wenfang Miao, Wei Liu, Yang Li, Guangming Lu, Feiyun Wu, Shouju Wang
Summary: The study introduces a Generative Adversarial Network for Distribution Analysis (GANDA) to describe and generate intratumoral quantum dots distribution, offering a new approach to investigate NPs distribution and guide nanomedicine optimization.
JOURNAL OF CONTROLLED RELEASE
(2021)
Article
Green & Sustainable Science & Technology
Fan Zhang, Yingqi Zhang, Xinhong Zhang
Summary: Artificial intelligence has a profound impact on meteorology research, and deep learning, as an AI method, greatly improves the accuracy of weather forecasting. A deep learning model called MDPGAN is proposed in this paper, which introduces a differential privacy framework to reduce the risk of identifying real data. The MDPGAN model can generate synthetic weather data with similar statistical characteristics to real data, meeting the requirements of data augmentation and desensitization.
JOURNAL OF CLEANER PRODUCTION
(2023)
Article
Engineering, Multidisciplinary
Shuai Ma, Jianfeng Cui, Chin-Ling Chen, Xuhui Chen, Ying Ma
Summary: The study proposes ECG Deep Convolution Generative Adversarial Networks (ECG-DCGAN) to expand the arrhythmia dataset and solve the problem of data imbalance. Experimental results show that the proposed classification method significantly improves the accuracy of arrhythmia diagnosis.
Article
Computer Science, Information Systems
Wanliang Wang, Hangyao Tu, Jiacheng Chen, Fei Wu
Summary: Recently, there has been significant progress in image-to-image translation in the literature. However, issues such as border distortion and color distortion continue to persist in existing methods. These methods often use multiple channels, which make it difficult to find the gradient in optimizer, resulting in unsatisfactory results. To address this problem, we propose the Three-Channel Generative Adversarial Network, which decomposes color images into three RGB color channels and utilizes single channel generators and dual discriminators for adversarial training. The algorithm also includes specific discriminators responsible for texture and structure discrimination, and a revised loss function to improve translation accuracy. Experimental results on various datasets demonstrate clear improvement over the pix2pix method in terms of quality and quantity.
MULTIMEDIA TOOLS AND APPLICATIONS
(2023)
Article
Environmental Sciences
Andrew Hennessy, Kenneth Clarke, Megan Lewis
Summary: The study explores the use of Generative Adversarial Networks to generate realistic synthetic hyperspectral vegetation data while maintaining class relationships, leading to improved classification accuracy and efficiency in handling vegetation spectral data.
Article
Environmental Sciences
Yiheng Cai, Fuxing Wan, Shinan Lang, Xiangbin Cui, Zijun Yao
Summary: Bed topography and roughness are important factors in ice-sheet analyses. Existing ice-penetrating radar measurements have data gaps in some areas and need interpolation. The new multi-branch deep learning method, MB_DeepBedMap, can generate more realistic bed topography with sub-kilometer roughness.
Article
Computer Science, Artificial Intelligence
Shufei Zhang, Kaizhu Huang, Zhuang Qian, Rui Zhang, Amir Hussain
Summary: SimpleGAN is proposed in this study to address the issues of unstable training and poor generation in GANs. By learning and utilizing an additional latent space with simple low-dimensional distributions, SimpleGAN tackles the difficulties in measuring the divergence between highly complex distributions in high-dimensional space from a different perspective.
NEURAL COMPUTING & APPLICATIONS
(2021)
Article
Computer Science, Artificial Intelligence
Hsin-Yi Chen, Szu-Hao Huang
Summary: Applying the privacy-preserving generative adversarial imitation network (PPGAIN) algorithm allows for establishing a profitable financial trading strategy while protecting the privacy of individuals. The generated trading behavior sequences are difficult to classify as imitations of specific individuals.
Article
Geochemistry & Geophysics
Mingqiu Mao, Huajun Wang, Peng Nie, Shipeng Xiao, Ruijie Wu
Summary: This article presents a generative adversarial network architecture combined with the U-Net network that incorporates a self-attention mechanism to strengthen the correlation between seismic data, aiming to improve the network's reconstruction capacity on the continuity of seismic signals. The intelligent denoising of seismic data enabled by this network enhances labor efficiency compared to traditional approaches and shows strong generalization and robustness.
IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING
(2022)
Article
Computer Science, Artificial Intelligence
Saksham Jain, Gautam Seth, Arpit Paruthi, Umang Soni, Girish Kumar
Summary: The study introduces a method of data augmentation using Generative Adversarial Networks, which significantly improves the performance of Convolutional Neural Networks in surface defect classification tasks. Training with synthetic images leads to better classification results.
JOURNAL OF INTELLIGENT MANUFACTURING
(2022)
Article
Biology
Ruichen Rong, Shuang Jiang, Lin Xu, Guanghua Xiao, Yang Xie, Dajiang J. Liu, Qiwei Li, Xiaowei Zhan
Summary: MB-GAN is a novel simulation framework designed using generative adversarial networks to generate microbiome data that are indistinguishable from real data. Compared to traditional methods, MB-GAN does not require explicit statistical modeling assumptions and is easily applicable.
Article
Computer Science, Artificial Intelligence
Qianxi Zhao, Liu Yang, Nengchao Lyu
Summary: Excessive stress leads to degraded driving performance and increases the risk of road accidents. This paper proposes a real-time driver stress detection method using deep learning and generative adversarial networks, focusing on analyzing pupillary response data. By establishing different models and utilizing data augmentation, the accuracy of stress detection and the recognition rate of minority categories are improved.
EXPERT SYSTEMS WITH APPLICATIONS
(2024)
Article
Computer Science, Artificial Intelligence
Abdurrahman Ocal, Lale Ozbakir
Summary: Generative adversarial networks (GANs) are a key generative network model for generating fake samples from real ones. In particular, the DCGAN variant plays a significant role in improving GAN performance with its convolutional architecture. The proposed Supervised DCGAN (SDCGAN) method allows for creating a supervised network structure when using multi-category datasets.
Article
Geosciences, Multidisciplinary
Honggeun Jo, Michael J. Pyrcz
Summary: Modeling semivariograms to characterize spatial continuity is subjective and noisy due to experimental variations. The proposed ASMC method uses deep learning to automate semivariogram modeling, improving objectivity and utilization of spatial data for a wide range of spatial modeling projects. The CNN-based approach successfully learns spatial characteristics and predicts semivariogram parameters with high accuracy, demonstrating the effectiveness of the machine learning workflow.
MATHEMATICAL GEOSCIENCES
(2022)
Article
Energy & Fuels
Jose J. Salazar, Lean Garland, Jesus Ochoa, Michael J. Pyrcz
Summary: This study proposes a new method that takes into account spatial autocorrelation in machine learning and designs a fair train-test split. By applying the semivariogram model and modified rejection sampling, the method generates a test set with similar prediction difficulty as the planned real-world use of the model. The method outperforms other approaches in several empirical analyses and provides spatial aware sets ready for predictive machine learning problems.
JOURNAL OF PETROLEUM SCIENCE AND ENGINEERING
(2022)
Article
Geosciences, Multidisciplinary
Mahmood Shakiba, Larry W. Lake, Julia F. W. J. Gale, Michael Pyrcz
Summary: This work presents novel multiscale spatial data analytics using Ripley's K-function to study the arrangement of one-dimensional fractures. The statistical significance of the calculated Ripley's K-function is used to classify fracture spatial arrangements. Characterizations of fracture arrangements as a function of length scale and position are performed. A simulation technique is introduced to reconstruct spatial arrangements and generate fracture realizations similar to the observed fractures. Synthetic and field-measured fracture datasets are used for testing and demonstration. These methods can be applied to fracture datasets observed in various geological settings.
JOURNAL OF STRUCTURAL GEOLOGY
(2022)
Article
Energy & Fuels
Peixi Zhu, Shayan Tavassoli, Jenny Ryu, Michael Pyrcz, Matthew T. Balhoff
Summary: Injection of gel systems can effectively reduce CO2 leakage in storage reservoirs, and the treatment success is highly dependent on the selection of appropriate operating conditions for specific reservoir properties.
INTERNATIONAL JOURNAL OF OIL GAS AND COAL TECHNOLOGY
(2022)
Article
Multidisciplinary Sciences
Javier E. Santos, Michael J. Pyrcz, Masa Prodanovi
Summary: Digital rock images are computational representations that capture the geometrical complexity of systems present ubiquitously in nature. In recent years, their use has become widespread due to the increasing availability of repositories, and open-source physics simulators and analysis tools. This article presents a dataset of 3D binary geometries in a standardized format that represent a wide variety of geological and engineering systems.
Article
Energy & Fuels
Eduardo Maldonado-Cruz, Michael J. Pyrcz
Summary: Numerical models are crucial for forecasting subsurface fluid flow response to support optimal decision-making in developing subsurface resources. Though current methods focus on prediction accuracy and minimizing error, significant uncertainty necessitates considering the entire uncertainty distribution. New workflow integrates machine learning-based surrogate flow models to predict subsurface responses efficiently and accurately.
JOURNAL OF PETROLEUM SCIENCE AND ENGINEERING
(2022)
Article
Computer Science, Software Engineering
Javier E. Santos, Alex Gigliotti, Abhishek Bihani, Christopher Landry, Marc A. Hesse, Michael J. Pyrcz, Masa Prodanovic
Summary: MPLBM-UT is a lattice-Boltzmann library specifically designed for single- and two-phase flow simulations in porous media. It provides a comprehensive set of tools and interfaces that allow users to perform simulations and visualize results with ease.
Article
Energy & Fuels
Wendi Liu, Michael J. Pyrcz
Summary: Spatial anomaly detection is crucial for subsurface resource exploration and environmental remediation. However, existing methods often overlook spatial context. We propose an ensemble anomaly detection method that integrates domain expertise, spatial continuity, and scale of interest to effectively identify local anomalous regions.
ENERGY EXPLORATION & EXPLOITATION
(2023)
Article
Geochemistry & Geophysics
Honggeun Jo, Yongchae Cho, Michael Pyrcz, Hewei Tang, Pengcheng Fu
Summary: Estimating porosity models from seismic data is challenging due to low signal-to-noise ratio and insufficient resolution. In this paper, a machine learning-based workflow is proposed to convert seismic data into porosity models. The workflow uses a residual U-Net++ architecture to estimate porosity models from multiple poststack seismic volumes. Experimental results show that the method achieves high accuracy and robustness in porosity estimation, but further research and improvements are needed.
Article
Computer Science, Interdisciplinary Applications
Jose Luis Hernandez-Mejia, Jesse Pisel, Honggeun Jo, Michael J. Pyrcz
Summary: Waterflooding is the most commonly used secondary method for oil recovery. We propose a novel approach called physics-constrained dynamic time warping (PCDTW) to evaluate the influence of water injection wells on oil production wells. This method can accurately determine the lag time between water injection and oil production response, providing valuable insights into reservoir connectivity and heterogeneity between paired wells.
COMPUTATIONAL GEOSCIENCES
(2023)
Article
Computer Science, Interdisciplinary Applications
Lei Liu, Masa Prodanovic, Michael J. Pyrcz
Summary: This study explores the impact of geostatistical nonstationarity on the prediction performance of CNNs in subsurface studies. The results show that various forms of geostatistical nonstationarity can affect the accuracy of CNN predictions. Therefore, it is important to consider the impact of geostatistical nonstationarity when using CNNs for subsurface data analysis.
COMPUTATIONAL GEOSCIENCES
(2023)
Article
Geochemistry & Geophysics
Wen Pan, Carlos Torres-Verdin, Ian J. Duncan, Michael J. Pyrcz
Summary: Well-log interpretation is crucial for estimating rock properties and supporting reservoir development. However, measurement errors and variability can affect the accuracy of estimations. To improve multiwell rock-property estimation, we propose discriminative adversarial (DA) and linear constraint models for well-log normalization and estimation. The DA model considers the joint distribution of well logs and rock properties, while the linear constraint model uses an ensemble of linear predictions to constrain normalization and estimation. The results show that the DA model outperforms other models by reducing the mean-squared error of permeability prediction by 20%-50%.
Article
Energy & Fuels
Wendi Liu, Michael J. Pyrcz
Summary: Production forecast based on historical data is essential but traditional methods are computationally intense or ignore subsurface geometries. Analytical data-driven models have limitations in capturing physics constraints while machine learning-based models may overfit due to sparse training data. We propose a grid-free, physics-informed graph neural network (PI-GNN) for accurate and interpretable production forecasting.
GEOENERGY SCIENCE AND ENGINEERING
(2023)