Related references
Note: Only part of the references are listed.
Article
Computer Science, Artificial Intelligence
Chunwei Tian et al.
Summary: This paper proposes a multi-stage image denoising CNN with wavelet transform, using dynamic convolution, wavelet transform and enhancement, and residual block to improve denoising performance. Experimental results show that the proposed method outperforms popular denoising methods.
PATTERN RECOGNITION
(2023)
Article
Geochemistry & Geophysics
Tian Renfei et al.
Summary: In this study, a wavelet denoising of multi-angle (WDMA) method is proposed for processing seismic survey data. The method utilizes wavelet decomposition and average weighting coefficient calculation to achieve denoising and stacking simultaneously. The results demonstrate that the proposed method can effectively reduce noise and obtain high-quality seismic data.
Article
Geosciences, Multidisciplinary
Xiangqi Wang et al.
Summary: This study used machine learning algorithms and SHAP analysis to predict and explain the impact of site conditions on seismic ground motion. The results showed that machine learning algorithms performed significantly better than traditional approaches, with XGBoost algorithm being the most effective. The explanation provided by the SHAP analysis enhanced the interpretability of the study. The combination of machine learning and SHAP analysis is expected to improve site amplification assessment in seismic hazard analysis.
FRONTIERS IN EARTH SCIENCE
(2023)
Article
Energy & Fuels
Ning Ma et al.
Summary: This paper proposes a method based on the Harris hawks optimization algorithm and long short-term memory recurrent neural networks to improve the prediction accuracy of the remaining useful life of supercapacitors, thus ensuring the safe operation of the entire energy storage system.
Article
Nuclear Science & Technology
E. Aymerich et al.
Summary: This work develops a data driven model based on Convolutional Neural Networks (CNNs) for real-time detection of disruptive events at JET. The model learns relevant spatiotemporal information directly from 1D plasma profiles, eliminating the need for manual feature extraction. Radiation profiles from both bolometer horizontal and vertical cameras are used as inputs to discriminate between core radiation and outboard radiation phenomena. The model architecture includes two separate CNNs trained to focus on events with different timescales, and their outputs are combined using a logic OR function to trigger the disruption alarm.
FUSION ENGINEERING AND DESIGN
(2023)
Article
Engineering, Manufacturing
Wei Li et al.
Summary: This study proposes a new data-driven approach for accurately predicting the remaining useful life (RUL) of tools in metal forming processes. By utilizing deep learning algorithms, this method can classify the wear states of workpiece surfaces with high accuracy and accurately predict the RUL values.
ADVANCES IN MANUFACTURING
(2023)
Article
Mathematical & Computational Biology
Norah Saleh Alghamdi et al.
Summary: This paper discusses the use of voice recognition analysis to provide insights into neurological illnesses for early diagnosis. Two deep learning models are introduced for detecting voice pathologies caused by neurological illnesses. The study found that the 2D CNN model performed better in terms of training and validation loss, and outperformed previous studies. The results suggest that a model trained on handcrafted features is more effective in speech processing.
COMPUTATIONAL AND MATHEMATICAL METHODS IN MEDICINE
(2022)
Article
Computer Science, Information Systems
Saumya Borwankar et al.
Summary: This paper aims to diagnose lung pathology using deep learning with a Convolutional Neural Network (CNN) and proposes a novel approach for processing audio samples. The research shows that the proposed method performs better in diagnosing lung sounds and does not require specific wheezing or crackling sounds.
MULTIMEDIA TOOLS AND APPLICATIONS
(2022)
Article
Computer Science, Artificial Intelligence
Gabriel Coelho et al.
Summary: This paper introduces three deep AutoEncoders for unsupervised Acoustic Anomaly Detection tasks and demonstrates their competitiveness through extensive experiments. In collaboration with an automotive company, a prototype of an in-vehicle intelligent system was developed, yielding interesting results.
NEURAL COMPUTING & APPLICATIONS
(2022)
Article
Oncology
Reid Shaw et al.
Summary: In this study, machine learning algorithms were used to identify the best combination of biomarkers for predicting ovarian malignancy in women with a pelvic mass. By combining multiple classifiers with the measurement of three distinct biomarkers, valuable preoperative diagnostic predictions can be made.
Article
Geochemistry & Geophysics
Laura Laurenti et al.
Summary: This study improves the prediction of labquakes using deep learning methods and expands the research scope. The autoregressive model is introduced to predict fault zone shear stress, and forecasts beyond a single seismic cycle are explored. The results show that these methods outperform existing models and have important implications for earthquake forecasting.
EARTH AND PLANETARY SCIENCE LETTERS
(2022)
Article
Geochemistry & Geophysics
Matteo Ravasi et al.
Summary: Multidimensional deconvolution is a powerful technique for removing overburden effects in seismic data, and the proposed stochastic gradient descent algorithm improves its stability and convergence.
IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING
(2022)
Article
Engineering, Electrical & Electronic
Md. Shah Fahad et al.
Summary: This paper proposes a speaker-adaptive DNN-HMM SER system that utilizes both MFCC and epoch-based features, incorporating speaker adaptation using feature space maximum likelihood linear regression technique. Experimental results emphasize the importance of speaker adaptation for SER systems and the complementary nature of MFCC and epoch-based features in emotion recognition.
CIRCUITS SYSTEMS AND SIGNAL PROCESSING
(2021)
Article
Multidisciplinary Sciences
Paul A. Johnson et al.
Summary: Through Google's ML competition platform Kaggle, utilizing knowledge and creativity from the machine learning community, a competition was conducted using laboratory earthquake data for forecasting. Unexpected strategies were employed by over 4,500 participating teams during the competition.
PROCEEDINGS OF THE NATIONAL ACADEMY OF SCIENCES OF THE UNITED STATES OF AMERICA
(2021)
Article
Computer Science, Software Engineering
Angelos Chatzimparmpas et al.
Summary: Ensemble methods like stacking are effective in improving predictive performance in machine learning by combining heterogeneous base models. However, creating a stack of models from scratch can be challenging due to the vast space of available solutions and different factors to consider such as data instances, features, and algorithms.
IEEE TRANSACTIONS ON VISUALIZATION AND COMPUTER GRAPHICS
(2021)
Article
Multidisciplinary Sciences
Kun Wang et al.
Summary: By leveraging data-driven machine learning, significant progress has been made in predicting fault-slip behavior in laboratory experiments, however, sparse seismic event data presents a challenge in training models. Utilizing transfer learning with numerical simulations, a model is trained to predict fault-slip behavior in experiments and improves predictions by further training with limited data.
NATURE COMMUNICATIONS
(2021)
Article
Energy & Fuels
Wanhu Zhang et al.
Summary: This study proposed a gradient boosting decision tree (GBDT) model based on the selection of input variables for predicting the cooling load of commercial buildings. By conducting correlation analysis and Pearson analysis to determine the optimal input variable combination, using ensemble learning technology to build the model, and applying the grid search method to determine the hyperparameter combination, the prediction accuracy of the model was improved.
Article
Green & Sustainable Science & Technology
Chen Cao et al.
Summary: The study explores the spatiotemporal distribution of earthquake activity and utilizes LSTM networks for pattern recognition of a synthetical earthquake catalog. The results show that applying LSTM in spatiotemporal series prediction has the potential for applications in the study of earthquake mechanics and forecasting major earthquake events.
Article
Energy & Fuels
Kai-Heng Sang et al.
Summary: This paper presents a new machine learning reservoir prediction method based on virtual sample generation, which improves prediction accuracy by generating virtual samples in a high-dimensional hypersphere space that are more consistent with the original data characteristics. Virtual sample screening and model iterative optimization are used to ensure the rationality of virtual samples.
Article
Engineering, Electrical & Electronic
Runwen Hu et al.
Summary: The letter proposes a novel CNN-based prediction approach trained on ImageNet for enhancing image prediction performance. Experimental results demonstrate that this method can effectively utilize surrounding pixels to improve prediction performance.
IEEE SIGNAL PROCESSING LETTERS
(2021)
Article
Computer Science, Interdisciplinary Applications
Giulia Baldazzi et al.
COMPUTER METHODS AND PROGRAMS IN BIOMEDICINE
(2020)
Article
Environmental Sciences
Ruhollah Taghizadeh-Mehrjardi et al.
Article
Computer Science, Information Systems
Sima Sahu et al.
MULTIMEDIA TOOLS AND APPLICATIONS
(2019)
Article
Physics, Multidisciplinary
Jian Cao et al.
PHYSICA A-STATISTICAL MECHANICS AND ITS APPLICATIONS
(2019)
Article
Chemistry, Multidisciplinary
Juan J. Noda et al.
APPLIED SCIENCES-BASEL
(2019)
Article
Geochemistry & Geophysics
Yuhan Sui et al.
Article
Geosciences, Multidisciplinary
Claudia Hulbert et al.
Article
Computer Science, Artificial Intelligence
Alaa Sagheer et al.
Article
Geochemistry & Geophysics
Thilo Wrona et al.
Article
Computer Science, Information Systems
Yanli Li et al.
Article
Geosciences, Multidisciplinary
K. M. Asim et al.
Article
Geochemistry & Geophysics
Tao Xie et al.
Article
Engineering, Electrical & Electronic
Merabi Mirel et al.
Article
Engineering, Electrical & Electronic
Qianwei Zhou et al.
IEEE SIGNAL PROCESSING LETTERS
(2012)
