Article
Energy & Fuels
Marco Ippolito, John Ferguson, Fred Jenson
Summary: Facies classification from well logs is crucial in seismic interpretation and reservoir characterization. Machine learning methods, specifically combining supervised and unsupervised learning, can improve accuracy in this task. The proposed multi-agent approach reduces bias during training and allows for a more accurate modeling of well log signals.
JOURNAL OF PETROLEUM SCIENCE AND ENGINEERING
(2021)
Article
Energy & Fuels
Cenk Temizel, Uchenna Odi, Karthik Balaji, Hakki Aydin, Javier E. Santos
Summary: Lithology and facies type are important factors in drilling operations and reservoir production behavior. Geological formations are complex and require sophisticated evaluation methods. Image classification and artificial intelligence algorithms can be used to accurately classify and identify geological formations, with supervised algorithms yielding more accurate results.
Article
Computer Science, Artificial Intelligence
Feng Liu, Guangquan Zhang, Jie Lu
Summary: This article introduces a shared-fuzzy-equivalence-relation neural network (SFERNN) for addressing the multisource heterogeneous UDA problem, which optimizes parameters by minimizing cross-entropy loss and distributional discrepancy between source and target domains. Experimental results demonstrate that SFERNN outperforms existing single-source heterogeneous UDA methods on multiple real-world datasets.
IEEE TRANSACTIONS ON FUZZY SYSTEMS
(2021)
Article
Computer Science, Artificial Intelligence
Riddhish Bhalodia, Shireen Elhabian, Ladislav Kavan, Ross Whitaker
Summary: This paper proposes a self-supervised deep learning approach for discovering landmarks from images that can be directly used as a shape descriptor. The method circumvents segmentation and preprocessing, producing a usable shape descriptor using just 2D or 3D images. Additionally, two variants on the training loss function are proposed to integrate prior shape information into the model.
MEDICAL IMAGE ANALYSIS
(2021)
Article
Computer Science, Artificial Intelligence
Keqiuyin Li, Jie Lu, Hua Zuo, Guangquan Zhang
Summary: Unsupervised domain adaptation leverages knowledge from a labeled source domain to tackle a similar unlabeled target domain. Existing methods transfer knowledge from a single source domain, but this may be insufficient. This paper proposes a dynamic classifier alignment (DCA) method for multi-source domain adaptation, which aligns classifiers driven by multi-view features.
IEEE TRANSACTIONS ON KNOWLEDGE AND DATA ENGINEERING
(2023)
Article
Meteorology & Atmospheric Sciences
Ghazal Farhani, Robert J. Sica, Mark Joseph Daley
Summary: The study uses machine learning techniques to train machines to sort measurements before processing and successfully classify lidar profiles with an identification success rate above 95% using various supervised machine learning algorithms. The t-SNE method, an unsupervised algorithm, is also utilized to cluster lidar profiles into meaningful categories and identify anomalies such as stratospheric aerosol layers due to wildfires.
ATMOSPHERIC MEASUREMENT TECHNIQUES
(2021)
Article
Environmental Sciences
Zhu Liang, Changming Wang, Zhijie Duan, Hailiang Liu, Xiaoyang Liu, Kaleem Ullah Jan Khan
Summary: This study employed a hybrid model that utilized the advantages of both supervised and unsupervised learning, and through a two-stage modeling process, constructed a robust landslide prediction model with improved performance.
Review
Plant Sciences
Jun Yan, Xiangfeng Wang
Summary: Advances in high-throughput omics technologies have led to the era of big data in plant biology research. Machine learning plays a crucial role in plant systems biology due to its excellent performance and wide application in analyzing big data. However, supervised machine learning algorithms require a large number of labeled samples as training data to achieve optimal performance. In cases where obtaining sufficient labeled training data is impossible or expensive, unsupervised learning and semi-supervised learning paradigms are indispensable.
Article
Computer Science, Artificial Intelligence
Gulsen Akman, Bahadir Yorur, Ali Ihsan Boyaci, Ming-Chuan Chiu
Summary: This study proposes the use of unsupervised and supervised machine learning algorithms to categorize companies based on their innovation capabilities. The companies are divided into three groups: good, satisfactory, and unsatisfactory, in order to establish a comprehensive and reliable assessment procedure. The study utilizes unsupervised and supervised machine learning methods to address the innovation capability evaluation problem.
APPLIED SOFT COMPUTING
(2023)
Article
Computer Science, Information Systems
Lars Schmarje, Monty Santarossa, Simon-Martin Schroeder, Reinhard Koch
Summary: Current deep learning strategies in computer vision are highly dependent on labeled data, which may not be feasible for many real-world problems. Therefore, incorporating unlabeled data and addressing issues like class imbalance and robustness is crucial. Future research trends include scalability, decreasing supervision needs, and combining ideas from different methods for improved performance.
Article
Geochemistry & Geophysics
Zhaokui Li, Qiang Xu, Li Ma, Zhuoqun Fang, Yan Wang, Wenqiang He, Qian Du
Summary: This study proposes an unsupervised domain adaptation method based on supervised contrastive learning, which optimizes the classification task by enhancing the separability of data within domains. A novel domain similarity loss is introduced for the target domain and a sample selection strategy is designed to improve the model's discrimination towards target domain data. Experimental results demonstrate the superiority of the proposed method over existing domain adaptation methods.
IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING
(2023)
Article
Chemistry, Multidisciplinary
Bang Xiao, Chunyue Lu
Summary: Improving the quantity and quality of labeled data is an effective way to enhance the performance of deep neural networks in computer vision tasks. However, obtaining high-quality annotations in medical image analysis and processing is challenging due to the requirement of expert knowledge. To address this, we propose a new semi-supervised framework that combines semi-supervised classification with unsupervised deep clustering. By spreading label information to unlabeled data, our framework extracts semantic information and improves model robustness. Experimental results on benchmark medical image datasets demonstrate the effectiveness of our method, achieving superior performance compared to existing methods.
APPLIED SCIENCES-BASEL
(2023)
Article
Computer Science, Artificial Intelligence
Kunwoong Kim, Ilsang Ohn, Sara Kim, Yongdai Kim
Summary: This paper investigates the validity of existing surrogate fairness constraints in fair AI and proposes a new surrogate fairness constraint called SLIDE, which is computationally feasible and achieves a fast convergence rate.
Article
Medicine, General & Internal
Taimoor Shakeel Sheikh, Jee-Yeon Kim, Jaesool Shim, Migyung Cho
Summary: This study proposes an unsupervised deep learning model for whole-slide image diagnosis. By fusing holistic and local appearance features and utilizing multiple image descriptors for training, the model achieves higher performance and diagnostic accuracy.
Article
Environmental Sciences
Ben Evans, Anita Faul, Andrew Fleming, David G. Vaughan, J. Scott Hosking
Summary: Accurate quantification of iceberg populations is crucial for estimating Southern Ocean freshwater and heat balances, as well as shipping hazards. Existing automated monitoring methods lack generality. This study proposes an adaptive unsupervised classification approach based on SAR data, which can robustly identify icebergs in complex environments. The method achieved good results in the evaluation of the study area.
REMOTE SENSING OF ENVIRONMENT
(2023)