Article
Geochemistry & Geophysics
Caijun Ren, Xiangyu Wang, Jian Gao, Xiren Zhou, Huanhuan Chen
Summary: This study introduces a novel change detection framework utilizing Generative Adversarial Network (GAN) to generate better coregistered images, improving the performance of change detection algorithms. Experimental results demonstrate that this method is less sensitive to the issue of unregistered images and effectively utilizes deep learning structures.
IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING
(2021)
Article
Computer Science, Information Systems
Mahsa Soleimani, Ali Nazari, Mohsen Ebrahimi Moghaddam
Summary: DeepFake involves the use of deep learning and artificial intelligence techniques to produce or change video and image contents. It can lead to fake news, crimes, and affect facial recognition systems. This study presents a deep learning approach using the entire face and face patches to distinguish real/fake images, overcoming limitations like blurring and obstruction. Experimental results show that this approach performs better and weighing the patches improves accuracy.
MULTIMEDIA SYSTEMS
(2023)
Article
Meteorology & Atmospheric Sciences
Emily Vosper, Peter Watson, Lucy Harris, Andrew McRae, Raul Santos-Rodriguez, Laurence Aitchison, Dann Mitchell
Summary: Flooding caused by intense rainfall is the main cause of death and damages from tropical cyclones. With the increase of rainfall caused by tropical cyclones under anthropogenic climate change, accurately estimating extreme rainfall is essential for resilience efforts. High-resolution climate models perform better in capturing TC statistics, but they are computationally expensive. Downscaling can be used to predict high-resolution features from low-resolution models. In this study, three deep learning models are developed and evaluated for downscaling TC rainfall, and the Wasserstein Generative Adversarial Network performs the best overall.
JOURNAL OF GEOPHYSICAL RESEARCH-ATMOSPHERES
(2023)
Article
Geochemistry & Geophysics
Fan Meng, Tao Song, Danya Xu
Summary: This study uses a generative adversarial network (GAN) and satellite infrared images to predict rainfall in tropical cyclones, solving the challenge of low temporal resolution in microwave sensors for estimating rainfall in tropical cyclones. The experimental results show that the algorithm can effectively extract key features and has the potential for global application, providing important insights for real-time visualization of rainfall in tropical cyclones.
IEEE GEOSCIENCE AND REMOTE SENSING LETTERS
(2022)
Article
Meteorology & Atmospheric Sciences
Teng Long, Jiyang Fu, Biao Tong, Pakwai Chan, Yuncheng He
Summary: This study utilizes deep learning techniques to identify the center location of tropical cyclones (TCs) based on TC satellite cloud images. Comparing six deep learning models, the YOLOv4 model achieved the highest confidence score and demonstrated excellent performance in identifying multiple TC locations and tracking individual TCs.
INTERNATIONAL JOURNAL OF CLIMATOLOGY
(2022)
Article
Geochemistry & Geophysics
Jean-Christophe Burnel, Kilian Fatras, Remi Flamary, Nicolas Courty
Summary: This article discusses how to generate adversarial attack examples in the case of black-box neural networks, and experiments were conducted using a specific method to demonstrate its effectiveness in image generation and modification. A perceptual evaluation with human annotators was also conducted to assess the effectiveness of the proposed method.
IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING
(2022)
Article
Engineering, Electrical & Electronic
Somrita Chattopadhyay, Avinash C. Kak
Summary: Despite the variability in appearance of buildings worldwide, our proposed generative adversarial network-based segmentation framework has achieved notable results in automatic building detection, particularly with improvements in edge and reverse attention.
IEEE JOURNAL OF SELECTED TOPICS IN APPLIED EARTH OBSERVATIONS AND REMOTE SENSING
(2022)
Article
Engineering, Electrical & Electronic
Selim Surucu, Banu Diri
Summary: With the increasing number of satellites, there is a rise in Earth observation imagery. This study utilizes generative adversarial networks to obtain fake images from the EuroSAT dataset and creates a dataset consisting of 14 classes and 36,000 images. By employing transfer learning models and ensemble models, the classification accuracy reaches 91.55%.
JOURNAL OF ELECTRONIC IMAGING
(2023)
Article
Computer Science, Information Systems
Azelle Courtial, Guillaume Touya, Xiang Zhang
Summary: This article examines the potential of using deep learning, specifically generative adversarial networks (GAN), to generate generalised mountain road maps. The study demonstrates the potential of deep learning in generating generalised maps and explores the working principle of deep learning generalisation, comparison between supervised and unsupervised learning, and possible improvements.
INTERNATIONAL JOURNAL OF GEOGRAPHICAL INFORMATION SCIENCE
(2023)
Article
Radiology, Nuclear Medicine & Medical Imaging
Yaru Sun, Yunqi Li, Pengfei Wang, Dongzhi He, Zhiqiang Wang
Summary: This paper proposes a novel approach for gastric lesion segmentation using generative adversarial training, which utilizes a segmentation network and a discriminator for accurate segmentation. Experimental results show that this method outperforms existing models and meets the needs of clinical diagnosis and treatment.
JOURNAL OF DIGITAL IMAGING
(2022)
Article
Geochemistry & Geophysics
Yixiang Huang, Ming Wu, Jun Guo, Chuang Zhang, Mengqiu Xu
Summary: This letter proposes a correlation context-driven method for sea fog detection, which utilizes a two-stage superpixel-based fully convolutional network and a fully connected Conditional Random Field (CRF) to model the relationships between pixels. An attentive Generative Adversarial Network (GAN) is also implemented for image enhancement. The experimental results show that the proposed method achieves high accuracy in detecting small, broken bits, and weak contrast thin structures.
IEEE GEOSCIENCE AND REMOTE SENSING LETTERS
(2022)
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
Geochemistry & Geophysics
Xu Chen, Bangguo Yin, Songqiang Chen, Haifeng Li, Tian Xu
Summary: This study focuses on the multiscale translation of satellite images to maps using generative adversarial networks. Previous studies were limited to smaller scales, so the researchers proposed a series strategy of generators to address this limitation. Experimental results showed that this strategy can achieve higher quality multiscale map generation.
IEEE GEOSCIENCE AND REMOTE SENSING LETTERS
(2022)
Article
Geosciences, Multidisciplinary
Mingliang Liu, Tapan Mukerji
Summary: This paper introduces a method based on deep generative adversarial networks to generate high-resolution digital rock images and recover fine details. Experimental results demonstrate the effectiveness of this method and its potential to better characterize heterogeneous porous media and predict pore-scale flow and petrophysical properties.
GEOPHYSICAL RESEARCH LETTERS
(2022)
Article
Environmental Sciences
Wei Zhong, Deyuan Zhang, Yuan Sun, Qian Wang
Summary: This study developed a CatBoost-based intelligent tropical cyclone intensity-detecting model using FY-2F and FY-2G CTBT data and CMA-BST best-track data. Compared to previous studies using pure CNN models, the CatBoost-based model exhibited better skills in detecting TC intensity, with an RMSE of 3.74 m/s.
Article
Biochemical Research Methods
Jianmin Wang, Yanyi Chu, Jiashun Mao, Hyeon-Nae Jeon, Haiyan Jin, Amir Zeb, Yuil Jang, Kwang-Hwi Cho, Tao Song, Kyoung Tai No
Summary: This study constructs a dataset for protein-protein interaction (PPI) targeted drug-likeness and proposes a deep molecular generative framework to generate novel drug-like molecules based on the features of seed compounds. The results show that the generated molecules have better PPI-targeted drug-likeness and drug-likeness, and the model performs comparably to other state-of-the-art molecule generation models.
BRIEFINGS IN BIOINFORMATICS
(2022)
Article
Computer Science, Theory & Methods
Fan Meng, Danya Xu, Tao Song
Summary: This study develops a real-time tropical cyclone wave height forecasting system using an adaptive time-frequency decomposition method and a mixture of deep learning methods. The system performs well in predicting wave heights caused by tropical cyclones and achieves excellent results in recent super typhoon events.
FUTURE GENERATION COMPUTER SYSTEMS-THE INTERNATIONAL JOURNAL OF ESCIENCE
(2022)
Article
Biochemical Research Methods
Tao Song, Xudong Zhang, Mao Ding, Alfonso Rodriguez-Paton, Shudong Wang, Gan Wang
Summary: In this study, DeepFusion, a deep learning based multi-scale feature fusion method, is proposed for predicting drug-target interactions. Experimental results show that DeepFusion achieves good prediction performance on different sub-datasets.
Article
Biochemical Research Methods
Xue Li, Peifu Han, Wenqi Chen, Changnan Gao, Shuang Wang, Tao Song, Muyuan Niu, Alfonso Rodriguez-Paton
Summary: This study proposes a protein-protein interaction (PPI) prediction model called multi-scale architecture residual network for PPIs (MARPPI) that utilizes dual-channel and multi-feature methods. The model leverages Res2vec to obtain residue association information and uses various descriptors to capture amino acid composition and order information, physicochemical properties, and information entropy. MARPPI achieves high accuracy rates ranging from 91.80% to 99.01% on different datasets, outperforming other advanced methods. The results also demonstrate the model's ability to detect hidden interactions and predict cross-species interactions.
BRIEFINGS IN BIOINFORMATICS
(2023)
Article
Biology
Xun Wang, Lulu Wang, Shuang Wang, Yongqi Ren, Wenqi Chen, Xue Li, Peifu Han, Tao Song
Summary: Molecular toxicity prediction is crucial for drug discovery and human health. Existing machine learning models for toxicity prediction do not fully utilize the 3D information of molecules, which can influence their toxicity. In this study, we propose QuantumTox, the first application of quantum chemistry in drug molecule toxicity prediction. Our model extracts quantum chemical information as 3D features and uses ensemble learning methods to improve accuracy and generalization. Experimental results demonstrate consistent outperformance compared to baseline models, even on small datasets.
COMPUTERS IN BIOLOGY AND MEDICINE
(2023)
Article
Environmental Sciences
Fan Meng, Yichen Yao, Zhibin Wang, Shiqiu Peng, Danya Xu, Tao Song
Summary: This study proposes a machine learning approach for probabilistic forecasting of tropical cyclone intensity. Previous studies cannot directly characterize the uncertainty in TC forecasting and suffer from computational effort issues. This study introduces a new method of evaluating the forecast without this uncertainty through the forecast distribution. The model outperforms current operational models and provides reliable probabilistic forecasts critical for disaster warnings.
ENVIRONMENTAL RESEARCH LETTERS
(2023)
Article
Computer Science, Artificial Intelligence
Fan Meng, Kunlin Yang, Yichen Yao, Zhibin Wang, Tao Song
Summary: Tropical cyclones are extreme disasters with significant impact on human beings, and forecasting their intensity has been a challenging task. Deep learning-based intensity forecasting has shown potential to outperform traditional methods, but inherent uncertainty in weather forecasting needs to be quantified for decision-making and risk assessment. This study proposes an intelligent system, PTCIF, based on deep learning to quantify uncertainty using multimodal meteorological data, the first study to assess uncertainty of tropical cyclones using a deep learning approach. Probabilistic forecasts are made for the intensity of 6-24 hours, showing comparable performance to weather forecast centers in terms of deterministic forecasts. Reliable prediction intervals and probabilistic forecasts are obtained, which are crucial for disaster warning and expected to complement operational models.
INTERNATIONAL JOURNAL OF INTELLIGENT SYSTEMS
(2023)
Article
Biochemical Research Methods
Tao Song, Yongqi Ren, Shuang Wang, Peifu Han, Lulu Wang, Xue Li, Alfonso Rodriguez-Paton
Summary: Deep learning has greatly improved and changed the process of de novo molecular design. The proposed DNMG model utilizes a deep generative adversarial network combined with transfer learning to consider the 3D spatial information and physicochemical properties of molecules, generating valid and novel drug-like ligands. The computational results demonstrate that the molecules generated by DNMG have better binding ability and physicochemical properties for target proteins.
Article
Genetics & Heredity
Linfang Jiao, Yongqi Ren, Lulu Wang, Changnan Gao, Shuang Wang, Tao Song
Summary: Advancements in single-cell sequencing research have revolutionized our understanding of cellular heterogeneity and functional diversity. However, scRNA-seq data analysis remains a computational challenge due to the high dimensionality and sparsity of the data, as well as the time-consuming and subjective nature of manual cell type identification.
FRONTIERS IN GENETICS
(2023)
Article
Biochemistry & Molecular Biology
Linfang Jiao, Gan Wang, Huanhuan Dai, Xue Li, Shuang Wang, Tao Song
Summary: Single-cell transcriptomics is advancing our understanding of complex tissues and cells, with scRNA-seq holding great potential for cell composition identification. However, manual annotation is time-consuming and unreliable for scRNA-seq data analysis. This paper introduces scTransSort, a cell-type annotation method based on scRNA-seq data and Transformer concept, which reduces data sparsity and computational complexity for cell type identification.
Article
Environmental Sciences
Tao Song, Jiarong Wang, Jidong Huo, Wei Wei, Runsheng Han, Danya Xu, Fan Meng
Summary: This study aims to develop a new deep learning algorithm, EEMD-LSTM, to accurately predict the significant wave height (SWH) of deep and distant ocean. The results show that the EEMD-LSTM model outperforms other comparative models in short-term and medium- and long-term SWH predictions, with RMSEs of 0.0204, 0.0279, 0.0452, 0.0941, and 0.1949 for the future 1, 3, 6, 12, and 18 h, respectively. It can serve as a rapid SWH prediction system to ensure navigation safety and has great practical significance and application value.
FRONTIERS IN MARINE SCIENCE
(2023)
Review
Geosciences, Multidisciplinary
Tao Song, Cong Pang, Boyang Hou, Guangxu Xu, Junyu Xue, Handan Sun, Fan Meng
Summary: The utilization and exploitation of marine resources by humans have contributed to the growth of marine research. With advancing technology, artificial intelligence (AI) approaches are being applied to maritime research, complementing traditional marine forecasting models and observation techniques. This article explores the application of AI in ocean observation, phenomena identification, and element forecasting.
FRONTIERS IN EARTH SCIENCE
(2023)
Article
Biochemical Research Methods
Shudong Wang, Chuanru Ren, Yulin Zhang, Yunyin Li, Shanchen Pang, Tao Song
Summary: In this study, a novel predictive model called RPCA$\Gamma $NR is proposed, which utilizes a new robust PCA framework based on $\gamma $-norm and $l_{2,1}$-norm regularization and an augmented Lagrange multiplier method to optimize it, thereby deriving the association scores for SM-miRNA. Through extensive evaluation, RPCA$\Gamma $NR outperforms existing models in terms of accuracy, efficiency, and robustness, significantly streamlining the process of determining SM-miRNA associations.
BRIEFINGS IN BIOINFORMATICS
(2023)
Article
Biotechnology & Applied Microbiology
Su-Kui Jin, Li-Na Xu, Yu-Jia Leng, Ming-Qiu Zhang, Qing-Qing Yang, Shui-Lian Wang, Shu-Wen Jia, Tao Song, Ruo-An Wang, Tao Tao, Qiao-Quan Liu, Xiu-Ling Cai, Ji-Ping Gao
Summary: In this study, the researchers identified a NAC transcription factor, OsNAC24, that regulates starch synthesis in rice. Through analysis of osnac24 mutants, it was found that OsNAC24 regulates starch synthesis by modulating the mRNA and protein levels of OsGBSSI and OsSBEI. Additionally, OsNAC24 interacts with another NAC transcription factor, OsNAP, to coactivate the expression of target genes. These findings highlight the important role of the OsNAC24-OsNAP complex in fine-tuning starch synthesis in rice endosperm.
PLANT BIOTECHNOLOGY JOURNAL
(2023)
Article
Computer Science, Artificial Intelligence
Handan Sun, Tao Song, Ying Li, Kunlin Yang, Danya Xu, Fan Meng
Summary: This study proposes a hybrid model based on ensemble empirical mode decomposition and Convolutional long short-term memory network to solve the non-smoothness problem in sea surface wind speed prediction. Experimental findings show that this model has the best prediction effect, and this advantage becomes increasingly evident as time increases.
APPLIED INTELLIGENCE
(2023)