检索结果-内蒙古大学图书馆

Accuracy of generative deep learning model for macular anatomy prediction from optical coherence tomography images in macular hole surgery

引用

SCIENTIFIC REPORTS 2024年第1期14卷 1-13页

作者： Kwon, Han Jo Heo, Jun Park, Su Hwan Park, Sung Who Byon, Iksoo Pusan Natl Univ Sch Med Biomed Res Inst Pusan Natl Univ HospDept Ophthalmol Gudeok Ro 179 Busan 49241 South Korea Pusan Natl Univ Yangsan Hosp Res Inst Convergence Biomed Sci & Technol Dept Ophthalmol Geumo Ro 20 Yangsan Si 50612 Gyeongsangnam D South Korea

This study aims to propose a generative deep learning model (GDLM) based on a variational autoencoder that predicts macular optical coherence tomography (OCT) images following full-thickness macular hole (FTMH) surgery and evaluate its clinical accuracy. Preoperative and 6-month postoperative swept-source OCT data were collected from 150 patients with successfully closed FTMH using 6 x 6 mm2 macular volume scan datasets. Randomly selected and augmented 120,000 training and 5000 validation pairs of OCT images were used to train the GDLM. We assessed the accuracy and F1 score of concordance for neurosensory retinal areas, performed Bland-Altman analysis of foveolar height (FH) and mean foveal thickness (MFT), and predicted postoperative external limiting membrane (ELM) and ellipsoid zone (EZ) restoration accuracy between artificial intelligence (AI)-OCT and ground truth (GT)-OCT images. Accuracy and F1 scores were 94.7% and 0.891, respectively. Average FH (228.2 vs. 233.4 mu m, P = 0.587) and MFT (271.4 vs. 273.3 mu m, P = 0.819) were similar between AI- and GT-OCT images, within 30.0% differences of 95% limits of agreement. ELM and EZ recovery prediction accuracy was 88.0% and 92.0%, respectively. The proposed GDLM accurately predicted macular OCT images following FTMH surgery, aiding patient and surgeon understanding of postoperative macular features.

关键词： Deep learning Generative deep learning model Macular hole Optical coherence tomography variational autoencoder

来源：评论

学校读者我要写书评

暂无评论

3D Generative Model Latent Disentanglement via Local Eigenprojection

引用

COMPUTER GRAPHICS FORUM 2023年第6期42卷 e14793-e14793页

作者： Foti, Simone Koo, Bongjin Stoyanov, Danail Clarkson, Matthew J. UCL London England Univ Calif Santa Barbara Santa Barbara CA USA

Designing realistic digital humans is extremely complex. Most data-driven generative models used to simplify the creation of their underlying geometric shape do not offer control over the generation of local shape attributes. In this paper, we overcome this limitation by introducing a novel loss function grounded in spectral geometry and applicable to different neural-network-based generative models of 3D head and body meshes. Encouraging the latent variables of mesh variational autoencoders (VAEs) or generative adversarial networks (GANs) to follow the local eigenprojections of identity attributes, we improve latent disentanglement and properly decouple the attribute creation. Experimental results show that our local eigenprojection disentangled (LED) models not only offer improved disentanglement with respect to the state-of-the-art, but also maintain good generation capabilities with training times comparable to the vanilla implementations of the models. Our code and pre-trained models are available at .

关键词： disentanglement generative adversarial networks geometric deep learning variational autoencoder

来源：评论

学校读者我要写书评

暂无评论

variational Bayesian inference with complex geostatistical priors using inverse autoregressive flows

引用

COMPUTERS & GEOSCIENCES 2023年 171卷

作者： Levy, Shiran Laloy, Eric Linde, Niklas Univ Lausanne Inst Earth Sci Lausanne Switzerland Belgian Nucl Res Ctr Inst Environm Hlth & Safety Mol Belgium

We combine inverse autoregressive flows (IAF) and variational Bayesian inference (variational Bayes) in the context of geophysical inversion parameterized with deep generative models encoding complex priors. variational Bayes approximates the unnormalized posterior distribution parametrically within a given family of distributions by solving an optimization problem. Although prone to bias if the chosen family of distributions is too limited, it provides a computationally-efficient approach that scales well to high-dimensional inverse problems. To enhance the expressiveness of the variational distribution, we explore its combination with IAFs that allow samples from a simple base distribution to be pushed forward through a series of invertible transformations onto an approximate posterior. The IAF is learned by maximizing the lower bound of the evidence (marginal likelihood), which is equivalent to minimizing the Kullback-Leibler divergence between the approximation and the target posterior distribution. In our examples, we use either a deep generative adversar-ial network (GAN) or a variational autoencoder (VAE) to parameterize complex geostatistical priors. Although previous attempts to perform Gauss-Newton inversion in combination with GANs of the same architecture were proven unsuccessful, the trained IAF provides a good reconstruction of channelized subsurface models for both GAN-and VAE-based inversions using synthetic crosshole ground-penetrating-radar data. For the considered examples, the computational cost of our approach is seven times lower than for Markov chain Monte Carlo (MCMC) inversion. Furthermore, the VAE-based approximations in the latent space are in good agreement. The VAE-based inversion requires only one sample to estimate gradients with respect to the IAF parameters at each iteration, while the GAN-based inversions need more samples and the corresponding posterior approximation is less accurate.

关键词： Geophysical inversion Normalizing flows variational inference Generative adversarial network variational autoencoder

来源：评论

学校读者我要写书评

暂无评论

Deep generative learning for exploration in large electrochemical impedance dataset

引用

ENGINEERING APPLICATIONS OF ARTIFICIAL INTELLIGENCE 2023年第PartC期126卷

作者： Doonyapisut, Dulyawat Kim, Byeongkyu Kim, Jung Kyu Lee, Eunseok Chung, Chan-Hwa Sungkyunkwan Univ Sch Chem Engn Suwon 16419 South Korea Sungkyunkwan Univ Coll Comp & Informat Suwon 16419 South Korea

The fields of energy storage, photocatalysis, and sensors have undergone substantial technological advancements, which have led to the generation of vast amounts of data on electrochemical impedance (EIS). The interpretation of large amounts of EIS data is a challenging task since the analysis of EIS data requires multiple steps to get a suitable equivalent circuit. Recently, some progress has been made in the machine learning (ML) model for EIS classification. However, most of the ML models are performed as a "black box" model, which provides only the classification result and lacks physical descriptor representation. Here, we apply variational autoencoders (VAE) to EIS data analysis, which includes classification, parameter prediction, and the visualization of physical descriptors. The VAE model performed well in the classification task, with an accuracy of 82.0%-92.4%. In the prediction task, VAE shows a high R-squared value on the Randles circuit. Additionally, the VAE model can map physical descriptors to the latent space, allowing the latent space to transform into a property space, which plays an important role in the optimization and exploration of novel materials research.

关键词： Electrochemical impedance spectroscopy Equivalent circuit model variational autoencoder Impedance optimization Impedance visualization

来源：评论

学校读者我要写书评

暂无评论

Deep generative fuel design in low data regimes via multi-objective imitation

引用

CHEMICAL ENGINEERING SCIENCE 2023年第1期274卷

作者： Liu, Yifan Liu, Runze Duan, Jinyu Wang, Li Zhang, Xiangwen Li, Guozhu Tianjin Univ Sch Chem Engn & Technol Key Lab Green Chem Technol Minist Educ Tianjin 300072 Peoples R China Collaborat Innovat Ctr Chem Sci & Engn Tianjin Tianjin 300072 Peoples R China Haihe Lab Sustainable Chem Transformat Tianjin 300192 Peoples R China

Commercial fuel discovery faces a constantly decreasing return of investment due to due to increasingly tight environmental criteria and reducing potential uses for each new fuel. In this paper, a deep generative model, termed Latent Interspace Generative Adversarial Network with a Domain of Stacking (LIGANDS), has been established to screen desired fuel molecules in the large chemical space without setting design rules manually. A variational autoencoder, a generative adversarial network and a stacking model are well integrated in LIGANDS through model convergence. Given only the structures of 255 typical highenergy???density fuels in low data regimes, LIGANDS generated 3461 new fuel molecules with similar property distribution and improved energy performance as the qualified candidates of next-generation fuels. To expand and enrich the fuel-relevant chemical space with innovative molecular entities on demand, in-depth multi-objective imitation on the key properties of target fuel is realized by LIGANDS through optimizing generative molecular structures and their distribution. ?? 2023 Elsevier Ltd. All rights reserved.

关键词： Deep learning Generative model Fuel variational autoencoder Generative adversarial network Ensemble learning

来源：评论

学校读者我要写书评

暂无评论

Data-Driven Virtual Material Analysis and Synthesis for Solid Electrolyte Interphases

引用

ADVANCED ENERGY MATERIALS 2023年第40期13卷 2301985-2301985页

作者： Rajagopal, Deepalaxmi Koeppe, Arnd Esmaeilpour, Meysam Selzer, Michael Wenzel, Wolfgang Stein, Helge Nestler, Britta Karlsruhe Inst Technol KIT Inst Appl Mat IAM MMS Str Forum 7 D-76131 Karlsruhe Germany Karlsruhe Inst Technol KIT Inst Nanotechnol INT Hermann Von Helmholtz Pl 1 D-76344 Eggenstein Leopoldshafen Germany Helmholtz Inst Ulm Lise Meitner Str 16 D-89081 Ulm Germany

Solid electrolyte interphases (SEIs) form as reduction products at the electrodes and strongly affect battery performance and safety. Because SEI formation poses a highly nonlinear, complex multi-physics problem over various lengths and time scales, traditional modeling approaches struggle to characterize SEI evolution solely with existing physical properties. To improve the characterization of SEIs, it proposes a data-driven strategy for a virtual material design that learns to represent and characterize SEI formation with physical and data-driven properties from kinetic Monte Carlo simulations. A variational autoencoder with a property regressor learns data-driven properties, which represent SEI configurations and correlate with physical target properties. This new neural network design encodes the high-dimensional structural and reaction spaces into a lower-dimensional latent space, while the property regressor orders the latent space by physical target properties. The model achieves high correlation scores between target and predicted properties from latent representations, thereby proving that the data-driven properties enrich the expressiveness of SEI characterizations. A recent study has developed a parallel regressor variational autoencoder tool that can automate the characterization of solid electrolyte interphase (SEI) in batteries. This tool can synthesize and explain SEI configurations by analyzing data-driven properties encoded with physical properties. The study also investigates the impact of reaction barriers on SEI ***

关键词： kinetic Monte Carlo simulations solid electrolyte interphases variational autoencoder virtual material design

来源：评论

学校读者我要写书评

暂无评论

Few-Shot User-Adaptable Radar-Based Breath Signal Sensing

引用

SENSORS 2023年第2期23卷 804-804页

作者： Mauro, Gianfranco Diez, Maria De Carlos Ott, Julius Servadei, Lorenzo Cuellar, Manuel P. Morales-Santos, Diego P. Infineon Technol AG Am Campeon 1 15 D-85579 Neubiberg Germany Univ Granada Dept Elect & Comp Technol Ave de Fuente Nueva s n Granada 18071 Spain Tech Univ Munich Dept Elect & Comp Engn Arcis str 21 D-80333 Munich Germany Univ Granada Dept Comp Sci & Artificial Intelligence C Pdta Daniel Saucedo Aranda s n Granada 18015 Spain

Vital signs estimation provides valuable information about an individual's overall health status. Gathering such information usually requires wearable devices or privacy-invasive settings. In this work, we propose a radar-based user-adaptable solution for respiratory signal prediction while sitting at an office desk. Such an approach leads to a contact-free, privacy-friendly, and easily adaptable system with little reference training data. Data from 24 subjects are preprocessed to extract respiration information using a 60 GHz frequency-modulated continuous wave radar. With few training examples, episodic optimization-based learning allows for generalization to new individuals. Episodically, a convolutional variational autoencoder learns how to map the processed radar data to a reference signal, generating a constrained latent space to the central respiration frequency. Moreover, autocorrelation over recorded radar data time assesses the information corruption due to subject motions. The model learning procedure and breathing prediction are adjusted by exploiting the motion corruption level. Thanks to the episodic acquired knowledge, the model requires an adaptation time of less than one and two seconds for one to five training examples, respectively. The suggested approach represents a novel, quickly adaptable, non-contact alternative for office settings with little user motion.

关键词： vital sign sensing respiration signal artificial neural networks meta-learning radar FMCW few-shot learning autocorrelation variational autoencoder signal processing

来源：评论

学校读者我要写书评

暂无评论

Disentangling Forced Trends in the North Atlantic Jet From Natural Variability Using Deep Learning

引用

JOURNAL OF GEOPHYSICAL RESEARCH-ATMOSPHERES 2024年第14期129卷

作者： Hermoso, Alejandro Schemm, Sebastian Swiss Fed Inst Technol Inst Atmospher & Climate Sci Zurich Switzerland

Regional weather variability and extremes over Europe are strongly linked to variations in the North Atlantic jet stream, especially during the winter season. Projections of the evolution of the North Atlantic jet are essential for estimating the regional impacts of climate change. Therefore, separating forced trends in the North Atlantic jet from its natural variability is an extremely relevant task. Here, a deep learning based method, the Latent Linear Adjustment autoencoder (LLAE), is used for this purpose on an ensemble of fully-coupled climate simulations. The LLAE is based on a variational autoencoder and an additional linear component. The model uses detrended temperature and geopotential to predict the component of the zonal wind associated with natural variability. The residual between this prediction and the original wind field is interpreted as the forced component of the jet. The method is first tested for the geostrophic wind for which the forced trend can be obtained analytically from the difference between geostrophic wind computed from detrended and full geopotential. Despite the large variability of the total trends, the LLAE is shown to be effective in extracting the forced component of the trend for each individual ensemble member in both geostrophic and full wind fields. The LLAE-derived forced trend shows an increase in the upper-level zonal wind speed along a southwest-northeast oriented band over the ocean and a jet extension toward Europe. These are common characteristics over different periods and show some similarities to the upper-level zonal wind speed trend obtained from the ERA5 reanalysis. The North Atlantic jet stream, which is a narrow and strong air current flowing from west to east, has a substantial influence on daily weather variability and temperature and precipitation extremes in Europe. It is therefore essential to investigate how the North Atlantic jet is being modified under climate change. Trends in the North Atlantic jet a

关键词： jet streams North Atlantic forced trends dynamical adjustment variational autoencoder

来源：评论

学校读者我要写书评

暂无评论

Hybrid Unsupervised Time-Series Anomaly Detection for Industrial IoT Based on Spatio-Temporal Feature 18th

Hybrid Unsupervised Time-Series Anomaly Detection for Indus...

引用

18th China Conference on Wireless Sensor Networks, CWSN 2024

作者： Dong, Yu Yao, Muyan Qi, Peng Tao, Dan School of Electronic and Information Engineering Beijing Jiaotong University Beijing100044 China

ISBN: (纸本)9789819621880

The rapid development of the Industrial Internet of Things (IIoT) has enabled the communication of numerous devices, resulting in the generation of massive time-series data. The goal of time-series anomaly detection algorithms is to quickly detect anomalous behavior and trigger warnings. We propose a Hybrid Unsupervised Time-series Anomaly Detection method (HUTAD) that fully leverages two different anomaly detection strategies: spatial feature extraction by a variational autoencoder and temporal feature extraction by a variational autoencoder and a recurrent feature extraction LSTM structure with predictive capability. Through a dynamic fusion module based on the self-attention mechanism, we integrate the two methods to obtain the output, analyze the spatio-temporal patterns of the data, and detect potential anomalies. We incorporate adversarial training in our model to further improve its accuracy and robustness. Finally, we use three widely-used large-scale datasets to validate our model. The experimental results show that our method outperforms the baseline methods in F1-Score. © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2025.

关键词： Adversarial training Anomaly detection Industrial loT Self-attention variational autoencoder

来源：评论

学校读者我要写书评

暂无评论

variational graph auto-encoders for miRNA-disease association prediction

引用

METHODS 2021年 192卷 25-34页

作者： Ding, Yulian Tian, Li-Ping Lei, Xiujuan Liao, Bo Wu, Fang-Xiang Univ Saskatchewan Div Biomed Engn Saskatoon SK S7N 5A9 Canada Beijing Wuzi Univ Sch Informat Beijing 101125 Peoples R China Shaanxi Normal Univ Sch Comp Sci Xian 710119 Peoples R China Hainan Normal Univ Sch Math & Stat Haikou 571158 Hainan Peoples R China Univ Saskatchewan Dept Mech Engn Saskatoon SK S7N 5A9 Canada Univ Saskatchewan Dept Comp Sci Saskatoon SK S7N 5A9 Canada

Cumulative experimental studies have demonstrated the critical roles of microRNAs (miRNAs) in the diverse fundamental and important biological processes, and in the development of numerous complex human diseases. Thus, exploring the relationships between miRNAs and diseases is helpful with understanding the mechanisms, the detection, diagnosis, and treatment of complex diseases. As the identification of miRNA-disease associations via traditional biological experiments is time-consuming and expensive, an effective computational prediction method is appealing. In this study, we present a deep learning framework with variational graph auto-encoder for miRNA-disease association prediction (VGAE-MDA). VGAE-MDA first gets the representations of miRNAs and diseases from the heterogeneous networks constructed by miRNA-miRNA similarity, disease-disease similarity, and known miRNA-disease associations. Then, VGAE-MDA constructs two sub-networks: miRNA-based network and disease-based network. Combining the representations based on the heterogeneous network, two variational graph auto-encoders (VGAE) are deployed for calculating the miRNA-disease association scores from two subnetworks, respectively. Lastly, VGAE-MDA obtains the final predicted association score for a miRNA-disease pair by integrating the scores from these two trained networks. Unlike the previous model, the VGAE-MDA can mitigate the effect of noises from random selection of negative samples. Besides, the use of graph convolutional neural (GCN) network can naturally incorporate the node features from the graph structure while the variational autoencoder (VAE) makes use of latent variables to predict associations from the perspective of data distribution. The experimental results show that VGAE-MDA outperforms the state-of-the-art approaches in miRNA-disease association prediction. Besides, the effectiveness of our model has been further demonstrated by case studies.

关键词： miRNA-disease association Deep learning Graph convolutional network variational autoencoder

来源：评论

学校读者我要写书评

暂无评论

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案：

请选择收藏分类：

通借通还

建议与咨询 留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案： 新增检索档案 确定 取消

请选择收藏分类： 新增自定义分类 确定 取消

通借通还

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

请选择保存的检索档案：

请选择收藏分类：