Search

Medical image analysis has experienced different stages of development, especially with the emergence of deep learning. However, acquiring large-scale, high-quality labeled data to train a deep learning model takes time and effort. This paper proposes a semisupervised learning method for medical image segmentation using limited labeled data and large-scale unlabeled data. Inspired by the classic Generative Adversarial Network (GAN) and co-training strategy, we proposed a new Co-GAN framework to implement medical image segmentation. The proposed Co-GAN comprises two generators and one discriminator, in which two generators can provide mutual segmentation in formation to each other. Through adversarial training between generators and discriminators, Co-GAN achieved higher segmentation accuracy. The dataset used was the hippocampus in Medical Segmenlation Decathlon (MSD). There were foin· training data settings: 25 labeled slices/3,374 unlabeled slices; 50 labeled slices/3,349 unlabeled slices; 100 labeled slices/3,299 unlabeled slices; and 200 labeled slices/3,199 unlabeled slices. Three experiments were conducted for each data set: fully supervised learning based on a generator network using only labeled data (F-Generator), semi-supervised learning based on GAN (Semi-GAN), and semi-supervised learning based on Co-GAN. The experiments showed that Co-GAN improved the segmentation accuracy b) (1.9%, 2.6%, 1.1%, and 0.1%) compared to F-Generator and (2.2%, 0.8%, 0.5%, 0.7%) to Semi-GAN. [ABSTRACT FROM AUTHOR]

This paper introduces a method aimed at diagnosing the presence or absence of lesions by detecting anomalies in Chest X-ray images. The proposed approach is based on the PatchCore anomaly detection method, which extracts a feature vector containing location information of an image patch from normal image data and calculates the anomaly distance from the normal vector. However, applying PatchCore directly to medical image processing presents challenges due to the possibility of diseases occurring only in specific organs and the presence of image noise unrelated to lesions. In this study, we present an image alignment method that utilizes affine transformation parameter prediction to standardize already captured X-ray images into a specific composition. Additionally, we introduce a region-specific abnormality detection method that requires affine-transformed chest X-ray images. Furthermore, we propose a method to enhance application efficiency and performance through feature map hard masking. The experimental results demonstrate that our proposed approach achieved a maximum AUROC (Area Under the Receiver Operating Characteristic) of 0.774. Compared to a previous study conducted on the same dataset, our method shows a 6.9% higher performance and improved accuracy. [ABSTRACT FROM AUTHOR]

Electroencephalogram (EEG)-based emotion recognition plays an important role in brain-computer interface and mental health monitoring. The large amount of EEG data but the lacks of labeling, multi-feature attribute, and data uncertainty are the difficulties in its recognition problem. A multi-view semi-supervised Takagi–Sugeno–Kang (MV-SS-TSK) fuzzy system is developed for EEG emotion classification in this paper. In the learning of fuzzy system consequent, firstly, a novel joint learning of semi-supervised learning, sparse representation, and low-rank coding is developed for semi-supervised sparse consequent factor learning, which makes the consequent parameter learning as a pseudo-label-only optimization problem. In particular, to simplify fuzzy rules, the sparse constraint term ensures the consequent parameters to be sparse in rows. Secondly, the consequent factor learning in a single feature view is extended into the multi-view learning model. In particular, low-rank coding is considered in multi-view semi-supervised consequent parameter learning. The low-rank constraint on view-shared component of consequent factor is implemented to exploit global data structure. The sparse constraint on view-dependent component of consequent factor is implemented to retain the feature diversity representation. By minimizing the intersection between view-shared component and view-specific components for different views, MV-SS-TSK can take advantage of the intrinsic relationship between various features and capture the consistency from multi-view features. Experiments on the SEED dataset show the superior performance of the proposed fuzzy system. [ABSTRACT FROM AUTHOR]

In recent years, image semantic segmentation technology has developed rapidly, but image annotation usually requires a significant amount of human and financial resources, especially for remote sensing image annotation, which can be expensive and sometimes even unaffordable. To address this issue, this paper integrates the idea of curriculum learning into the self-training method and screens reliable pseudo-labels through computing image-level confidence, significantly reducing the confirmation error problem. Furthermore, the semi-supervised model in this paper combines implicit semantic enhancement with strong data augmentation, which can reduce the coupling between the teacher model and the student model's prediction distribution and enhance the model's robustness. Finally, the proposed semi-supervised method is experimentally verified using the ISPRS competition dataset and compared with existing state-of-the-art (SOTA) methods. Experimental results show that the proposed semi-supervised segmentation method achieves higher segmentation accuracy compared to self-training methods. Moreover, despite not using iterative training to simplify the training process, the proposed method still yields satisfactory segmentation results. [ABSTRACT FROM AUTHOR]

Background: Neoadjuvant therapy followed by surgery has become the standard of care for locally advanced esophageal squamous cell carcinoma (ESCC) and accurate pathological response assessment is critical to assess the therapeutic efficacy. However, it can be laborious and inconsistency between different observers may occur. Hence, we aim to develop an interpretable deep-learning model for efficient pathological response assessment following neoadjuvant therapy in ESCC. Methods: This retrospective study analyzed 337 ESCC resection specimens from 2020–2021 at the Pudong-Branch (Cohort 1) and 114 from 2021–2022 at the Puxi-Branch (External Cohort 2) of Fudan University Shanghai Cancer Center. Whole slide images (WSIs) from these two cohorts were generated using different scanning machines to test the ability of the model in handling color variations. Four pathologists independently assessed the pathological response. The senior pathologists annotated tumor beds and residual tumor percentages on WSIs to determine consensus labels. Furthermore, 1850 image patches were randomly extracted from Cohort 1 WSIs and binarily classified for tumor viability. A deep-learning model employing knowledge distillation was developed to automatically classify positive patches for each WSI and estimate the viable residual tumor percentages. Spatial heatmaps were output for model explanations and visualizations. Results: The approach achieved high concordance with pathologist consensus, with an R^2 of 0.8437, a RAcc_0.1 of 0.7586, a RAcc_0.3 of 0.9885, which were comparable to two senior pathologists (R^2 of 0.9202/0.9619, RAcc_0.1 of 8506/0.9425, RAcc_0.3 of 1.000/1.000) and surpassing two junior pathologists (R^2 of 0.5592/0.5474, RAcc_0.1 of 0.5287/0.5287, RAcc_0.3 of 0.9080/0.9310). Visualizations enabled the localization of residual viable tumor to augment microscopic assessment. Conclusion: This work illustrates deep learning's potential for assisting pathological response assessment. Spatial heatmaps and patch examples provide intuitive explanations of model predictions, engendering clinical trust and adoption (Code and data will be available at https://github.com/WinnieLaugh/ESCC%5fPercentage once the paper has been conditionally accepted). Integrating interpretable computational pathology could help enhance the efficiency and consistency of tumor response assessment and empower precise oncology treatment decisions. [ABSTRACT FROM AUTHOR]

Side-channel analysis (SCA) is a cryptanalytic technique that exploits physical leakages, such as power consumption or electromagnetic emanations, from cryptographic devices to extract secret keys used in cryptographic algorithms. Recent studies have shown that training SCA models with semi-supervised learning can effectively overcome the problem of few labeled power traces. However, the process of training SCA models using semi-supervised learning generates many pseudo-labels. The performance of the SCA model can be reduced by some of these pseudo-labels. To solve this issue, we propose the HWFilter method to improve semisupervised SCA. This method uses a Hamming Weight Pseudo-label Filter (HWPF) to filter the pseudo-labels generated by the semi-supervised SCA model, which enhances the model's performance. Furthermore, we introduce a normal distribution method for constructing the HWPF. In the normal distribution method, the Hamming weights (HWs) of power traces can be obtained from the normal distribution of power points. These HWs are filtered and combined into a HWPF. The HWFilter was tested using the ASCADv1 database and the AES_HD dataset. The experimental results demonstrate that the HWFilter method can significantly enhance the performance of semi-supervised SCA models. In the ASCADv1 database, the model with HWFilter requires only 33 power traces to recover the key. In the AES_HD dataset, the model with HWFilter outperforms the current best semi-supervised SCA model by 12%. [ABSTRACT FROM AUTHOR]

Infrared ship target segmentation is the important basis of infrared guided weapon in the sea-air context. Typically, accurate infrared ship target segmentation relies on a large number of pixel-level labels. However, it is difficult to obtain them. To this end, we present a method of Semi-supervised Infrared Ship Target Segmentation with Dual Branch (SeISTS-DB), which utilizes a small amount of labeled data and a large amount of unlabeled data to train model and improve segmentation performance. There are three main contributions. First, we design a target segmentation branch to generate the pseudo labels for unlabeled data. It consists of a dual learning network and a segmentation network. The dual learning network generates pseudo labels with weights for unlabeled data. The segmentation network is trained using both labeled data and unlabeled data with pseudo labels to achieve target segmentation of infrared ship, obtaining the preliminary segmentation results. Secondly, we introduce an error segmentation pixel correction branch, which contains a student network and a teacher network, to modify the pixel category error of the preliminary segmentation map. Finally, the outputs of the two branches are combined to obtain the final segmentation result. The SeISTS-DB is compared with other fully-supervised and semi-supervised methods on the infrared ship images dataset. Experimental results demonstrate that when the labeled data accounts for 1/8 of the training data, the mean Intersection over Union (mIou) is respectively improved by 15.35% and 6.19% at most. Besides, it is also compared with other methods on the public IRSTD-1k dataset, when the proportion of labeled images is 1/8, the mIoU is respectively improved by 11.76% at most compared to the state-of-the-art semi-supervised methods, demonstrating its effectiveness. [ABSTRACT FROM AUTHOR]

Non‐orthogonal multiple access (NOMA) is considered a key candidate technology for next‐generation wireless communication systems due to its high spectral efficiency and massive connectivity. Incorporating the concepts of multiple‐input‐multiple‐output (MIMO) into NOMA can further improve the system efficiency, but the hardware complexity increases. This study develops an energy‐efficient (EE) subchannel assignment framework for MIMO‐NOMA systems under the quality‐of‐service and interference constraints. This framework handles an energy‐efficient co‐training‐based semi‐supervised learning (EE‐CSL) algorithm, which utilizes a small portion of existing labeled data generated by numerical iterative algorithms for training. To improve the learning performance of the proposed EE‐CSL, initial assignment is performed by a many‐to‐one matching (MOM) algorithm. The MOM algorithm helps achieve a low complex solution. Simulation results illustrate that a lower computational complexity of the EE‐CSL algorithm helps significantly minimize the energy consumption in a network. Furthermore, the sum rate of NOMA outperforms conventional orthogonal multiple access. [ABSTRACT FROM AUTHOR]

The automation of Fault Detection and Diagnosis (FDD) is a central task for many industries today. A myriad of methods are in use, although the most recent leading contenders are data-driven approaches and especially Machine Learning (ML) methods. ML algorithms fall into two main categories: supervised and unsupervised methods, depending on whether or not the instances are labeled with the expected outputs. However, a new approach called Semi-Supervised Learning (SSL) has recently emerged that uses a few labeled instances together with other unlabeled instances for the training process. This new approach can significantly improve the accuracy of conventional ML models for industrial environments where labeled data are scarce. SSL has been tested as a promising solution over the past few years for several FDD problems, although there have been no systemic reviews of this sort of approach up until the present review. In this study, an attempt to organize the existing literature on SSL for FDD using the taxonomy of van Engelen & Hoos is reported. The most and the least frequently used SSL algorithms are identified and considered in terms of different fault detection tasks and their most common dataset structure. Moreover, a set of best practices are proposed in the conclusions of this work for implementation under real industrial conditions, so as to avoid some of the most common faults. • Presentation of a survey with 157 articles on Industrial Fault Detection and Diagnosis (FDD) using Semi-Supervised Learning (SSL). • Distribution of inductive SSL FDD solutions found in the review in terms of type of method. • Identification of factual standards of SSL algorithms for different FDD tasks and their most common dataset structure. • Identification of most common benchmark industrial datasets and its difference with laboratory datasets for SSL FDD. • A set of best practices are proposed for SSL algorithms' implementation under real industrial conditions in FDD tasks. [ABSTRACT FROM AUTHOR]

Artificial intelligence (AI) and machine learning (ML) approaches are increasingly being used in dementia research. However, several methodological challenges exist that may limit the insights we can obtain from high‐dimensional data and our ability to translate these findings into improved patient outcomes. To improve reproducibility and replicability, researchers should make their well‐documented code and modeling pipelines openly available. Data should also be shared where appropriate. To enhance the acceptability of models and AI‐enabled systems to users, researchers should prioritize interpretable methods that provide insights into how decisions are generated. Models should be developed using multiple, diverse datasets to improve robustness, generalizability, and reduce potentially harmful bias. To improve clarity and reproducibility, researchers should adhere to reporting guidelines that are co‐produced with multiple stakeholders. If these methodological challenges are overcome, AI and ML hold enormous promise for changing the landscape of dementia research and care. Highlights: Machine learning (ML) can improve diagnosis, prevention, and management of dementia.Inadequate reporting of ML procedures affects reproduction/replication of results.ML models built on unrepresentative datasets do not generalize to new datasets.Obligatory metrics for certain model structures and use cases have not been defined.Interpretability and trust in ML predictions are barriers to clinical translation. [ABSTRACT FROM AUTHOR]

How to accurately recognize the mental state of pilots is a focus in civil aviation safety. The mental state of pilots is closely related to their cognitive ability in piloting. Whether the cognitive ability meets the standard is related to flight safety. However, the pilot's working state is unique, which increases the difficulty of analyzing the pilot's mental state. In this work, we proposed a Convolutional Neural Network (CNN) that merges attention to classify the mental state of pilots through electroencephalography (EEG). Considering the individual differences in EEG, semi-supervised learning based on improvedK-Means is used in themodel training to improve the generalization ability of the model. We collected the EEG data of 12 pilot trainees during the simulated flight and compared the method in this paper with other methods on this data. The method in this paper achieved an accuracy of 86.29%, which is better than 4D-aNN and HCNN etc. Negative emotion will increase the probability of fatigue appearing, and emotion recognition is also meaningful during the flight. Then we conducted experiments on the public dataset SEED, and our method achieved an accuracy of 93.68%. In addition, we combine multiple parameters to evaluate the results of the classification network on a more detailed level and propose a corresponding scoring mechanism to display the mental state of the pilots directly. [ABSTRACT FROM AUTHOR]

The Internet of Things (IoT) is crucial in developing next-generation high-speed railways (HSRs). HSR IoT enables intelligent diagnosis of trains using multi-sensor data, which is critical for maintaining high speeds and ensuring passenger safety. Graph neural network (GNN)-based methods have gained popularity in HSR IoT research due to the ability to represent the sensor network as intuitive graphs. However, labeling monitoring data in the HSR scenario takes time and effort. To address this challenge, we propose a semi-supervised graph-level representation learning approach called MIM-Graph, which uses mutual information maximization to learn from a large amount of unlabeled data. First, the multi-sensor data is converted into association graphs based on their spatial topology. The unsupervised encoder is trained using global–local mutual maximization. The teacher–student framework transfers knowledge from the unsupervised encoder learned to the supervised encoder, which is trained using a small amount of labeled data. As a result, the supervised encoder learns distinguishable representations for intelligent diagnosis of HSR. We evaluate the proposed method using CWRU dataset and data from HSR Bogie test platform, and the experimental results demonstrate the effectiveness and superiority of MIM-Graph. • The fault diagnosis for the HSR bogie bearings under multisensor condition is transformed into a graph-level classification problem. • A semi-supervised method called MIM-Graph is proposed to maximize mutual information from two perspectives under limited labeled data. • MIM-graph utilizes two GNN-based encoders and is trained using a teacher–student framework. • The effectiveness and superiority of the proposed MIM-Graph are verificated by comparing with the state-of-art methods. [ABSTRACT FROM AUTHOR]

The study is devoted to the problem of label quality assurance and labeling cost minimization. It is important to have high quality labels in order to build efficient supervised machine learning pipelines. The aforementioned labels have proved to be costly to acquire; however, one can work with weaker instances of supervision that may be helpful to create a reliably set of ground truth labels for classifiers. The authors present a semi-supervised approach using pretrained language models to increase supervision information quality from general-purpose as well domain-specific document sets where weak labels and unlabeled data is present. [ABSTRACT FROM AUTHOR]

Quantifying the effect of mutations in the BRCA1 gene is useful for understanding their clinical consequences on breast cancer. Machine learning models can be applied to predict the landscape of protein variant effects that might not be always accessible by experiments. In this work, we propose a simple semi-supervised learning method using a Gaussian mixture model to predict ∼90% of the unlabeled missense variants of the BRCA1 gene collected from the ClinVar database. High-quality embeddings are used as a feature of the protein sequences, extracted using the latest pre-trained transformer-based language protein model. A statistical test show that the protein embeddings are effective and robust for predicting pathogenicity. Further, the lower representations of the features are then fed into the semi-supervised model. The prediction performance of the model only for the labeled testing data achieves an AUC score and an accuracy of 79.27% and 71.58%, respectively. Using our defined pathogenic probability score, we find that ∼94% of variants in our unlabeled dataset are well-separated into either benign or pathogenic classes according to that scoring. Our scores obtain a moderate Spearman rank correlation with the results of established unsupervised variant effect models. Finally, our approach can potentially be developed for more accurate and biologically reliable predictions of the variant effects. [ABSTRACT FROM AUTHOR]

Object counting presents a significant challenge for applications demanding high levels of automation. In recently ears, semi-supervised counting methods have been explored to alleviate the burden of data annotation. Despite considerable progress, the state-of-the-art methods treat pixels with varying levels of difficulty equally during each training iteration, resulting in an inflexible consistency training that can lead to high supervision loss at the onset of training and potentially cause network collapse. To address this limitation, we propose an uncertainty-adaptive consistency training algorithm for semi-supervised object counting networks in this paper. Our approach leverages image colorization to construct colorization uncertainty, which can measure the difficulty of object pixels and background pixels. To ensure superior uncertainty estimation, we then enhance this colorization uncertainty by using density features to guide the learning of colorization. Furthermore, an uncertainty-adaptive consistency training (UACT) algorithm is introduced for the consistency training of density maps. By adaptively adjusting the uncertainty threshold, we prioritize harder pixels for selection, gradually incorporating easier ones into the consistency training. The effectiveness o f t he UACT is evaluated through extensive experiments on two datasets. Experimental results demonstrate that our method outperforms state-of-the-art semi-supervised counting methods. [ABSTRACT FROM AUTHOR]

Robot-assisted airway intubation application needs high accuracy in locating targets and organs. Two vital landmarks, nostrils and glottis, can be detected during the intubation to accommodate the stages of nasal intubation. Automated landmark detection can provide accurate localization and quantitative evaluation. The Detection Transformer (DeTR) leads object detectors to a new paradigm with long-range dependence. However, current DeTR requires long iterations to converge, and does not perform well in detecting small objects. This paper proposes a transformer-based landmark detection solution with deformable DeTR and the semantic-aligned-matching module for detecting landmarks in robot-assisted intubation. The semantics aligner can effectively align the semantics of object queries and image features in the same embedding space using the most discriminative features. To evaluate the performance of our solution, we utilize a publicly accessible glottis dataset and automatically annotate a nostril detection dataset. The experimental results demonstrate our competitive performance in detection accuracy. Our code can be accessible at https://github.com/ConorLTH/airway-intubation_landmarks-detection. [ABSTRACT FROM AUTHOR]

In the realm of modern diagnostic technology, video capsule endoscopy (VCE) is a standout for its high efficacy and non-invasive nature in diagnosing various gastrointestinal (GI) conditions, including obscure bleeding. Importantly, for the successful diagnosis and treatment of these conditions, accurate recognition of bleeding regions in VCE images is crucial. While deep learning-based methods have emerged as powerful tools for the automated analysis of VCE images, they often demand large training datasets with comprehensive annotations. Acquiring these labeled datasets tends to be time-consuming, costly, and requires significant domain expertise. To mitigate this issue, we have embraced a semi-supervised learning (SSL) approach for the bleeding regions segmentation within VCE. By adopting the 'Mean Teacher' method, we construct a student U-Net equipped with an scSE attention block, alongside a teacher model of the same architecture. These models' parameters are alternately updated throughout the training process. We use the Kvasir-Capsule dataset for our experiments, which encompasses various GI bleeding conditions. Notably, we develop the segmentation annotations for this dataset ourselves. The findings from our experiments endorse the efficacy of the SSL-based segmentation strategy, demonstrating its capacity to reduce reliance on large volumes of annotations for model training, without compromising on the accuracy of identification. [ABSTRACT FROM AUTHOR]

With the development of Internet technology, multimodal data have become the main data resource in the information age. Multimodal learning mode, as an education and teaching mode developed on multimodal data technology, provides more convenience for multimedia teaching. However, many challenges persist in its actual development and application. Currently, the multimodal learning model is susceptible to classroom noise, lack of teaching information, and other factors, which adversely affect multimodal data collection, teaching application, and achievement output. Thus, this paper optimizes the multimodal learning model in the open learning environment and takes 120 engineering students from a university in Guangxi Province as the research object. First, a sequential modal extraction method is proposed by constructing a multimodal probability generation model and then the data are modeled. Semi-supervised learning is then achieved by analyzing and combining the supervised and unsupervised learning processes. Finally, the knowledge base technology with information fusion characteristics is applied to the multimodal teaching mode. This teaching mode has been proven to improve students' learning ability and learning achievement and teachers' teaching effectiveness. [ABSTRACT FROM AUTHOR]

Looking for local activities and events is a common task for most users during travel or daily life. Events are usually announced on the event organizers' website or spread by posting on social networks such as Facebook Event or Facebook Fanpages. Integrating all these activities/events allows us to explore the city and understand its dynamics. In this article, we study the problem of event extraction, including event title recognition, venue extraction, and relationship coupling. Although distant supervision is a common technique for generating annotated training data, how to choose proper seed entities depends on the nature of the entities to be extracted, and the automatic labeling strategy adopted. To improve the performance, we proposed model-based distant supervision for event title recognition and Point Of Interest (POI) extraction, which reached 0.565 and 0.536 F1, respectively. Meanwhile, we conduct sequential pattern mining from Facebook event posts to determine the event venue and start/end date when multiple addresses/POIs or temporal expressions are recognized in a message. Overall, the average F1 of the proposed model in event extraction is 0.620. [ABSTRACT FROM AUTHOR]

Copyright of Journal of Harbin Institute of Technology. Social Sciences Edition / Haerbin Gongye Daxue Xuebao. Shehui Kexue Ban is the property of Harbin Institute of Technology and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

Out of the estimated few trillion galaxies, only around a million have been detected through radio frequencies, and only a tiny fraction, approximately a thousand, have been manually classified. We have addressed this disparity between labeled and unlabeled images of radio galaxies by employing a semi-supervised learning approach to classify them into the known Fanaroff-Riley Type I (FRI) and Type II (FRII) categories. A Group Equivariant Convolutional Neural Network (G-CNN) was used as an encoder of the state-of-the-art self-supervised methods SimCLR (A Simple Framework for Contrastive Learning of Visual Representations) and BYOL (Bootstrap Your Own Latent). The G-CNN preserves the equivariance for the Euclidean Group E(2), enabling it to effectively learn the representation of globally oriented feature maps. After representation learning, we trained a fully-connected classifier and fine-tuned the trained encoder with labeled data. Our findings demonstrate that our semi-supervised approach outperforms existing state-of-the-art methods across several metrics, including cluster quality, convergence rate, accuracy, precision, recall, and the F1-score. Moreover, statistical significance testing via a t-test revealed that our method surpasses the performance of a fully supervised G-CNN. This study emphasizes the importance of semi-supervised learning in radio galaxy classification, where labeled data are still scarce, but the prospects for discovery are immense. [ABSTRACT FROM AUTHOR]

The Graph Partitioning Problem (GPP) is a classical combinatorial optimization problem that has been extensively researched. In recent years, many methods for solving the GPP have been proposed, which can be divided into direct partitioning approaches and iterative improvement approaches. Direct partitioning approaches directly give a complete partition of the input graph. Iterative improvement approaches require further adjustment of feasible solutions based on the result of direct partitioning approaches in order to obtain a better performance. In this paper, we summarize the recent trends in algorithms and applications for GPP. In addition, we propose a graph partitioning algorithm based on semi-supervised learning in combination with graph filtering methods. [ABSTRACT FROM AUTHOR]

Over recent decades, there has been a growing interest in the use of semisupervised clustering. Compared to the supervised or unsupervised clustering methods for solving different real-life problems, the review of relevant articles shows that semisupervised clustering methods are more powerful, and even a small amount of supervised information can significantly improve the results of unsupervised methods. One popular method for incorporating partial supervised information is the use of labeled data. In this study, a semi-supervised clustering algorithm called ConvexClust is proposed. The proposed method improves data clustering using a geometric view borrowed from the Lune concept in the connectivity index and 10% of labeled data. Use of labeled data and formation of a convex hull are the beginning steps toward clustering. Next, labeling of non-labeled data and updating of the convex hull in an iterative process are the next steps. Evaluations of three UCI datasets and sixteen artificial datasets indicate that the proposed method outperforms other semi-supervised and traditional clustering techniques. [ABSTRACT FROM AUTHOR]

To address the problem of low detection accuracy due to training noise caused by mislabeling when Tri-training for network intrusion detection (NID), we propose a Tri-training algorithm based on cross entropy and K-nearest neighbors (TCK) for network intrusion detection. The proposed algorithm uses cross-entropy to replace the classification error rate to better identify the difference between the practical and predicted distributions of the model and reduce the prediction bias of mislabeled data to unlabeled data; K-nearest neighbors are used to remove the mislabeled data and reduce the number of mislabeled data. In order to verify the effectiveness of the algorithm proposed in this paper, experiments were conducted on 12 UCI datasets and NSL-KDD network intrusion datasets, and four indexes including accuracy, recall, F-measure and precision were used for comparison. The experimental results revealed that the TCK has superior performance than the conventional Tri-training algorithms and the Tritraining algorithms using only cross-entropy or K-nearest neighbor strategy. [ABSTRACT FROM AUTHOR]

Traditional graph embedding methods only consider the pairwise relationship between fault data. But in practical applications, the relationship of high-dimensional fault data usually is multiple classes corresponding to multiple samples. Therefore, the hypergraph structure is introduced to fully portray the complex structural relationship of high-dimensional fault data. However, during the construction of the hypergraph, the hyperedge weight is usually set as the sum of the similarities between every two vertices contained within the hyperedge, and this "averaging effect" causes the relationship between data sample points with high similarity to be weakened, while the relationship between data sample points with low similarity to be strengthened. This phenomenon also leads to the hypergraph cannot accurately portray the relationship of high-dimensional data, which reduces the fault classification accuracy. To address this issue, a novel dimensionality reduction method named Semi-supervised Multi-Graph Joint Embedding (SMGJE) is proposed and applied to rotor fault diagnosis. SMGJE constructs simple graphs and hypergraphs with the same sample points and characterizes the structure of high-dimensional data in a multi-graph joint embedding. The edges of the simple graph are the direct description of the similarity between sample points so that SMGJE can overcome this "averaging effect" of the hypergraph. The effectiveness of the proposed method is verified by two different fault datasets. • To the best of the author's knowledge, we have applied hypergraph to the field of fault diagnosis for the first time. • This paper provides an effective approach to characterize the structural relationship of data in high-dimensional space. • The theory of semi-supervision is introduced to address the labeled samples limited issue. • A novel dimensionality reduction method named Semi-supervised Multi-Graph Joint Embedding (SMGJE) is proposed. [ABSTRACT FROM AUTHOR]

In industrial detection scenarios, achieving high accuracy typically relies on extensive labeled datasets, which are costly and time-consuming. This has motivated a shift towards semi-supervised learning (SSL), which leverages labeled and unlabeled data to improve learning efficiency and reduce annotation costs. This work proposes the unsupervised spectral clustering labeling (USCL) method to optimize SSL for industrial challenges like defect variability, rarity, and complex distributions. Integral to USCL, we employ the multi-task fusion self-supervised learning (MTSL) method to extract robust feature representations through multiple self-supervised tasks. Additionally, we introduce the Enhanced Spectral Clustering (ESC) method and a dynamic selecting function (DSF). ESC effectively integrates both local and global similarity matrices, improving clustering accuracy. The DSF maximally selects the most valuable instances for labeling, significantly enhancing the representativeness and diversity of the labeled data. USCL consistently improves various SSL methods compared to traditional instance selection methods. For example, it boosts Efficient Teacher by 5%, 6.6%, and 7.8% in mean Average Precision(mAP) on the Automotive Sealing Rings Defect Dataset, the Metallic Surface Defect Dataset, and the Printed Circuit Boards (PCB) Defect Dataset with 10% labeled data. Our work sets a new benchmark for SSL in industrial settings. [ABSTRACT FROM AUTHOR]

Computer vision-based precision spraying of herbicides presents a promising avenue for reducing herbicide input and weed control costs. Nonetheless, weed detection in wheat (Triticum aestivum L.) remains challenging. Developing an effective and reliable neural network for weed detection requires substantial labeled data for training. However, labeling data is time-consuming and labor-intensive. To address this challenge, the present study introduces semi-supervised learning (SSL) into the domain of weed detection in wheat. The performance of four SSL methods was thoroughly evaluated and compared with that of a fully supervised learning (FSL) method on a dataset with a limited amount of labeled images. Experimental results showed that the Fixmatch method, an SSL approach, outperformed the FSL method, exhibiting significantly higher accuracy (ACC) with a limited number of labeled images. The ACC of Fixmatch was 85.4%, which was 7.3% higher than the FSL method. In further analysis, the performance of models trained on a dataset containing 100, 200, 300, 400, 500, or 1000 labeled images per class was tested. Compared with FSL, SSL achieved the greatest improvement when the number of labels was 200. At the same time, Fixmatch achieved satisfactory performance, ACC, recall, and precision reached 94.8%, 94.8%, and 95.2%, respectively, and the F1 score was 95%. In summary, these results suggest that using the SSL method could yield a high-performing model when training with a limited number of labeled images, requiring less training costs and lower demands on manpower. • This paper introduces semi-supervised learning to detect weeds growing in wheat. • Semi-supervised learning outperformed fully supervised learning for weed detection. • Using semi-supervised methods reduced the requirement of training costs with fewer labeled data. [ABSTRACT FROM AUTHOR]

[Display omitted] Cognitive assessment plays a pivotal role in the early detection of cognitive impairment, particularly in the prevention and management of cognitive diseases such as Alzheimer's and Lewy body dementia. Large-scale screening relies heavily on cognitive assessment scales as primary tools, with some low sensitivity and others expensive. Despite significant progress in machine learning for cognitive function assessment, its application in this particular screening domain remains underexplored, often requiring labor-intensive expert annotations. This paper introduces a semi-supervised learning algorithm based on pseudo-label with putback (SS-PP), aiming to enhance model efficiency in predicting the high risk of cognitive impairment (HR-CI) by utilizing the distribution of unlabeled samples. The study involved 189 labeled samples and 215,078 unlabeled samples from real world. A semi-supervised classification algorithm was designed and evaluated by comparison with supervised methods composed by 14 traditional machine-learning methods and other advanced semi-supervised algorithms. The optimal SS-PP model, based on GBDT, achieved an AUC of 0.947. Comparative analysis with supervised learning models and semi-supervised methods demonstrated an average AUC improvement of 8% and state-of-art performance, repectively. This study pioneers the exploration of utilizing limited labeled data for HR-CI predictions and evaluates the benefits of incorporating physical examination data, holding significant implications for the development of cost-effective strategies in relevant healthcare domains. [ABSTRACT FROM AUTHOR]

Risk prediction plays a crucial role in planning for prevention, monitoring, and treatment. Electronic Health Records (EHRs) offer an expansive repository of temporal medical data encompassing both risk factors and outcome indicators essential for effective risk prediction. However, challenges emerge due to the lack of readily available gold-standard outcomes and the complex effects of various risk factors. Compounding these challenges are the false positives in diagnosis codes, and formidable task of pinpointing the onset timing in annotations. We develop a Se mi-supervised D ouble D eep Le arning Temporal R isk Prediction (SeDDLeR) algorithm based on extensive unlabeled longitudinal Electronic Health Records (EHR) data augmented by a limited set of gold standard labels on the binary status information indicating whether the clinical event of interest occurred during the follow-up period. The SeDDLeR algorithm calculates an individualized risk of developing future clinical events over time using each patient's baseline EHR features via the following steps: (1) construction of an initial EHR-derived surrogate as a proxy for the onset status; (2) deep learning calibration of the surrogate along gold-standard onset status; and (3) semi-supervised deep learning for risk prediction combining calibrated surrogates and gold-standard onset status. To account for missing onset time and heterogeneous follow-up, we introduce temporal kernel weighting. We devise a Gated Recurrent Units (GRUs) module to capture temporal characteristics. We subsequently assess our proposed SeDDLeR method in simulation studies and apply the method to the Massachusetts General Brigham (MGB) Biobank to predict type 2 diabetes (T2D) risk. SeDDLeR outperforms benchmark risk prediction methods, including Semi-parametric Transformation Model (STM) and DeepHit, with consistently best accuracy across experiments. SeDDLeR achieved the best C-statistics (0.815, SE 0.023; vs STM +.084, SE 0.030, P -value.004; vs DeepHit +.055, SE 0.027, P -value.024) and best average time-specific AUC (0.778, SE 0.022; vs STM + 0.059, SE 0.039, P -value.067; vs DeepHit + 0.168, SE 0.032, P -value < 0.001) in the MGB T2D study. SeDDLeR can train robust risk prediction models in both real-world EHR and synthetic datasets with minimal requirements of labeling event times. It holds the potential to be incorporated for future clinical trial recruitment or clinical decision-making. [ABSTRACT FROM AUTHOR]

Label distribution learning is a novel machine learning paradigm to deal with label ambiguity, which is the generalization of the traditional single-label learning and multi-label learning paradigms. Though label distribution learning has attracted a lot of attentions recently, data sets with label distributions rather than logical labels have always been scarce and most of the existing models put emphasis mainly on the supervised learning, neglecting the utilization of unlabeled samples. In this paper, we propose a semi-supervised Label Distribution Learning model with label Correlations and data Manifold exploration (sLDLCM). On one hand, sLDLCM is a semi-supervised extension of existing label distribution learning models, which can effectively make use of both labeled and unlabeled data for better capturing the underlying data properties; on the other hand, sLDLCM jointly estimates the label distributions of unlabeled samples and the other model variables. Besides, both label correlations and local data manifold are explored in sLDLCM. Extensive experiments are conducted on six real-world data sets including two facial expression, one movie rating, two bioinformatics and one visual sentiment analysis data sets. Comparative studies demonstrate that the proposed sLDLCM model achieves better performance the state-of-the-arts in terms of five evaluation metrics. [ABSTRACT FROM AUTHOR]

Label distribution learning is a novel machine learning paradigm to deal with label ambiguity, which is the generalization of the traditional single-label learning and multi-label learning paradigms. Though label distribution learning has attracted a lot of attentions recently, data sets with label distributions rather than logical labels have always been scarce and most of the existing models put emphasis mainly on the supervised learning, neglecting the utilization of unlabeled samples. In this paper, we propose a semi-supervised Label Distribution Learning model with label Correlations and data Manifold exploration (sLDLCM). On one hand, sLDLCM is a semi-supervised extension of existing label distribution learning models, which can effectively make use of both labeled and unlabeled data for better capturing the underlying data properties; on the other hand, sLDLCM jointly estimates the label distributions of unlabeled samples and the other model variables. Besides, both label correlations and local data manifold are explored in sLDLCM. Extensive experiments are conducted on six real-world data sets including two facial expression, one movie rating, two bioinformatics and one visual sentiment analysis data sets. Comparative studies demonstrate that the proposed sLDLCM model achieves better performance the state-of-the-arts in terms of five evaluation metrics. [ABSTRACT FROM AUTHOR]

Semi-supervised learning (SSL) is becoming the mainstream paradigm in medical image segmentation, which enables models to jointly leverage both annotated and unannotated images. Despite recent progress, several challenges significantly affect the reliability of SSL model: (1) the empirical distribution mismatch between labeled and unlabeled data results in reliable knowledge derived from limited labeled data being largely discarded; (2) the inherent cognitive biases of the model inevitably generate unreliable pseudo-labels, leading to confirmation bias. In this paper, we propose a reliable semi-supervised mutual learning framework (RSSML), which incorporates reliable knowledge utilization strategy into the mutual learning paradigm to address above challenges. Specifically, we first devise a recombination-and-recovery data augmentation strategy to mutually intermix labeled and unlabeled images. The recombined images are then fed into two subnets with entirely different network structures to promote the learning of common semantics among them. For labeled images, the prediction differences between subnets help identify regions prone to missegmentation. We devise a supervised discordance relearning (SDR) regularization to review these regions. Regarding unlabeled images, we propose a reliability-aware cross pseudo supervision (RCPS) regularization to evaluate the reliability of pseudo-labels from two subnets and select those reliable ones for cross supervision. Extensive experiments on both publicly available and clinically obtained medical image datasets demonstrate the superiority of our method against existing SSL methods. The code is available at: https://github.com/1KB0/RSSML. • Novel reliable semi-supervised mutual learning framework (RSSML) for medical image segmentation. • Novel recombination-and-recovery-based data augmentation addresses empirical distribution mismatch issue. • Novel supervised discordance relearning regularization leverages reliable knowledge obtained from limited labeled data. • Novel reliability-aware cross pseudo supervision regularization selects only the reliable pseudo-labels for cross supervision. [Display omitted] [ABSTRACT FROM AUTHOR]

The liquid loading of shale gas wells presents significant challenges, including a reduction in production capacity, an increase in production costs, and environmental degradation. It is of paramount importance to provide accurate and timely warnings of liquid loading in order to ensure the efficient production operations. Existing physical models often rely on assumptions regarding material properties and fluid behaviors. However, the establishment of precise physical models in complex real-world scenarios represents a significant challenge. Most existing data-driven methods rely on either supervised or unsupervised learning strategies. The former causes a considerable investment of time and human resources to generate labels, while the latter demands exceptionally high standards for data quality. These are challenging to achieve in actual production. Most studies on liquid loading adopt a dichotomous approach, whereby the presence or absence of liquid loading is determined. A paucity of studies has attempted to classify the severity of liquid loading. In this paper, a semi-supervised learning-based liquid loading severity classification model is proposed. The model integrates the benefits of supervised and unsupervised learning, enabling the utilization of the entire dataset with a minimal number of labels. The paper also presents a Dynamic Time Relationship Self-Attention (DTRSA) module, which enhances the model's anti-interference ability against other anomalies and improves the accuracy of liquid loading severity classification. In order to assess the effectiveness of the proposed model, experiments are conducted on a dataset of 219 shale gas wells in the southwestern region of Sichuan Province, China, within an oil and gas field. The model exhibits an overall classification accuracy of 98.46% in detecting liquid loading, with an earlier warning time than existing physical models. • Propose a liquid load detection method based on semi-supervised learning. • Developed DTRSA to enhance accuracy by addressing anomalies. • Validated in Sichuan's oil and gas fields against maintenance records. • Achieved 86.55% recognition, warned 1-3 days earlier than physical models. [ABSTRACT FROM AUTHOR]

• A novel PBNet is designed to predict labels and filter out the high-threshold data. • LDM is used to generate new data consistent with the distribution of original data. • Newly RandAugmentBA is employed to enhance the attention on color and anomaly points. • The utilization of LDM-generated data are improved by a new label fusion and dynamic thresholds strategies. • Our method can effectively solve the imbalanced-classes and have the best performance. Biliary Atresia (BA) is a congenital condition affecting the bile ducts in newborn babies worldwide, typically diagnosed by the color of infant stools in clinical practice. Due to the limited availability of pediatricians and the lack of effective diagnostic tools for parents, computer vision is devised as a non-invasive and convenient means for early detection of BA. However, the limited recognition accuracy of BA and the challenge of imbalanced classes in real-world scenarios restrict the generalizability of existing methods, posing a challenge to their decision-making performance. To address those challenges, we introduce a Latent Diffusion Biliary Atresia method (LDBA) to leverage imbalanced data to enhance the recognition of BA. Specifically, we propose a novel pre-trained Biliary Atresia network (PBNet) to predict labels and effectively filter out high-threshold data. Then, a new data augmentation method is proposed to focus attention on the color and anomaly points of images. Subsequently, a new label fusion method and dynamic thresholds strategy are devised to optimize the utilization of generated data by the Latent Diffusion Model (LDM). The entire process is trained in a semi-supervised manner. The experimental results indicate that our proposed LDBA outperforms competing methods with a superior performance of 78.43%. Our method can improve the accuracy and reliability of BA clinical diagnosis. [ABSTRACT FROM AUTHOR]

The goal of semi-supervised MRI image segmentation is to train high-accuracy segmentation models using a small amount of labeled data and a large amount of unlabeled data, thereby alleviating the burden of manual annotation by physicians. Self-training is a traditional semi-supervised learning method aimed at acquiring additional image features by generating pseudo-labels for unlabeled data. However, this method may lead to an unstable training process due to differences between samples. Additionally, the predictions of model may contain erroneous noise, leading to low-confidence pseudo-labels. To address these issues, we proposed a dual uncertainty-guided self-training framework (DUST) for semi-supervised MRI image segmentation. Specifically, the two uncertainties consist of sample-level uncertainty and pixel-level uncertainty. The sample-level uncertainty is intended to achieve a stable and smooth training process. The pixel-level uncertainty is intended to rectify pseudo-labels. We conducted a series of experiments on two public MRI image datasets, ACDC2017 and Prostate2018. Compared to the baseline, we improved the Dice scores by 5.0% and 4.0% over the two datasets, respectively. Furthermore, the experimental results indicate that the proposed method has certain advantages compared to the comparative methods. This validates the feasibility and applicability of our method. • An uncertainty-based self-training framework for MRI image segmentation. • Sample-level uncertainty is used to prioritize unlabeled data segmentation. • Pixel-level uncertainty is used to obtain high-quality pseudo labels. • Proposed method can effectively utilize unlabeled data to train high-accuracy models. [ABSTRACT FROM AUTHOR]

Semi-supervised learning (SSL) is a widely used model training paradigm that effectively utilizes a limited set of labeled data and a substantially larger pool of unlabeled data. Historically, the focus of SSL research has predominantly been on classification tasks, employing methods such as consistency regularization and pseudo-labeling. However, the direct application of these methods to regression tasks presents significant challenges, primarily due to the complexities associated with evaluating the reliability of pseudo-labels in a regression context. This paper introduces SimRegMatch, a novel semi-supervised regression (SSR) framework devised to overcome this specific challenge, by combining an uncertainty-based filtering mechanism with a similarity-based pseudo-label calibration approach. The former component is tasked with discerning which unlabeled examples possess pseudo-labels of sufficient reliability, achieved through the estimation of uncertainty levels. The latter component then refines these pseudo-labels by propagating information from labeled to unlabeled examples, thereby enhancing the overall quality of the pseudo-labels. The efficacy of SimRegMatch was rigorously tested through experiments conducted on the publicly available AgeDB dataset, which is centered around age prediction, as well as on a practical regression problem focused on the detection of interior noise levels in automobiles using accelerometer data. When benchmarked against current state-of-the-art methods in semi-supervised regression, SimRegMatch exhibited notable improvements in regression performance. Additionally, a series of ablation studies were carried out to dissect and understand the specific elements of the framework that were instrumental in achieving these performance enhancements. SimRegMatch addresses a pivotal issue in semi-supervised regression – the assessment of regression pseudo-label reliability – and substantially elevates model performance. By combining the strengths of uncertainty estimation and pseudo-label calibration, SimRegMatch emerges as a robust and versatile framework with significant potential for broad applicability in various SSR scenarios. A PyTorch implementation is publicly available at https://github.com/YongwonJo/SimRegMatch. • Uncertainty-based filtering is used to select examples with reliable pseudo-label. • Similarity-based pseudo-label fuse information from labeled and unlabeled examples. • Calibrated pseudo-labels using labeled data improve regression performance. [ABSTRACT FROM AUTHOR]

Neoadjuvant chemotherapy (NAC) is the primary treatment used to reduce the tumor size in early breast cancer. Patients who achieve a pathological complete response (pCR) after NAC treatment have a significantly higher five-year survival rate. However, accurately predicting whether patients could achieve pCR remains challenging due to the limited availability of manually annotated MRI data. This study develops a weakly and semi-supervised joint learning model that integrates multi-parametric MR images to predict pCR to NAC in breast cancer patients. First, the attention-based multi-instance learning model is designed to characterize the representation of multi-parametric MR images in a weakly supervised learning setting. The Mean-Teacher learning framework is then developed to locate tumor regions for extracting radiochemical parameters in a semi-supervised learning setting. Finally, all extracted MR imaging features are fused to predict pCR to NAC. Our experiments were conducted on a cohort of 442 patients with multi-parametric MR images and NAC outcomes. The results demonstrate that our proposed model, which leverages multi-parametric MRI data, provides the AUC value of over 0.85 in predicting pCR to NAC, outperforming other comparative methods. • Neoadjuvant chemotherapy' early prediction is essential in breast cancer treatment. • Our method uses weakly and semi-supervised learning to maximize dataset utilization. • Our model takes entire MRI as the inputs including both bilateral and unilateral MRI. • Our designed model greatly reduces the labor intensity involved in manual annotation. • Our method (AUC value of 0.85) is superior to comparative methods in pCR prediction. [ABSTRACT FROM AUTHOR]

It is a big challenge to identify personalized driver genes (PDGs) for understanding tumor heterogeneity of cancer individual patients. From the perspective of machine learning, identifying PDGs is an inherent class imbalance issue due to the fewer known driver genes than most passenger ones. However, existing machine learning based methods including unsupervised and supervised learning based methods ignore the importance of limited well-established cancer tissue specific driver genes(CSDGs) for this class imbalance issue. Here we converted the PDG prediction issue to a semi-supervised classification task and a novel method (namely PersonalizedGNN) was developed to identify PDGs by using graph attention neural network and label reuse strategy in personalized gene interaction network (PGIN). PersonalizedGNN effectively utilizes the structure information of PGIN and the limited well-established CSDG information for achieving promising performance. Using the breast cancer and lung cancer datasets from The Cancer Genome Atlas, we validated our method and compared it with other advanced methods. PersonalizedGNN showed outstanding potential in identifying cancer driver genes in terms of prediction precision. Furthermore, we could discover subtype-specific de novo cancer driver genes and in vitro cell-based assays for a novel driver gene FZD7 in lung squamous cell carcinoma cells further validated the PersonalizedGNN. In summary, PersonalizedGNN offers a new effective perspective for discovering PDGs by considering information of prior known CSDGs in PGIN which help researchers understand tumor heterogeneity of cancer individual patients. • A novel method (namely PersonalizedGNN) was developed for identifying personalized driver genes using graph attention neural network and label reuse strategy. • PersonalizedGNN effectively utilizes the structure information of PGIN and the limited well-established cancer tissue-specific driver gene information during the semi-supervised learning process. • The computational experiments showed the superiority of PersonalizedGNN compared to other existing methods. • In vitro experiments further validated a novel driver gene FZD7 in cell proliferation, migration, and cell cycle regulation in the lung squamous cell carcinoma (LUSC) cell line (H1703). [ABSTRACT FROM AUTHOR]

In recent years, artificial intelligence has been applied to 3D COVID-19 medical image diagnosis, which reduces detection costs and missed diagnosis rates with higher predictive accuracy, and diagnostic efficiency. However, the limited size and low quality of clinical 3D medical image samples have hindered the segmentation performance of 3D models. Therefore, we propose a 3D medical image segmentation model based on semi-supervised learning using co-training. Multi-view and multi-modal images are generated using spatial flipping and windowing techniques to enhance the spatial diversity of 3D image samples. A pseudo label generation module based on confidence-weights is employed to generate reliable pseudo labels for non-annotated data, thereby increasing the sample size and reducing overfitting. The proposed approach utilizes a three-stage training process: firstly, training a single network based on annotated data; secondly, incorporating non-annotated data to train a dual-modal network and generate pseudo labels; finally, jointly training six models in three dimensions using both annotated and pseudo labels generated from multi-view and multi-modal images, aiming to enhance segmentation accuracy and generalization performance. Additionally, a consistency regularization loss is applied to reduce noises and accelerate convergence of the training. Moreover, a heatmap visualization method is employed to focus on the attention of features at each stage of training, providing effective reference for clinical diagnosis. Experiments were conducted on an open dataset of 3D COVID-19 CT samples and a non-annotated dataset from TCIA, including 771 NIFTI-format CT images from 661 COVID-19 patients. The results of 5-fold cross-validation show that the proposed model achieves a segmentation accuracy of Dice=73.30 %, ASD=10.633, Sensitivity=63.00 %, and Specificity=99.60 %. Compared to various typical semi-supervised learning 3D segmentation models, it demonstrates better segmentation accuracy and generalization performance. • Technique of 3D flipping and CT windowing are employed to construct a multi-view, multi-modal sample dataset. • By a non-linear transformation, uncertainty is converted into confidence-weights togenerate pseudo-labels with higher credibility. • A universally replaceable and parallelCNNs is used for multi-model co-training to enhance segmentation performance. • Consistency regularization loss is applied to reduce noises and accelerate convergence of training. • Heatmap visualization is employed to focus on the attention of features at each stage of co-training. [ABSTRACT FROM AUTHOR]

Semi-supervised learning methods have achieved great success in cardiac image analysis. This paper presents an effective semi-supervised cardiac image segmentation framework (SCS-SLSP) that utilizes subjective logic (SL) theory and the subset prototype generation (SPG) module. Specifically, based on the student–teacher model, SL theory generates soft labels with uncertainty for unlabeled data, distinguishing in-distribution pixels and out-of-distribution (uncertain) pixels located near class feature boundaries. The SPG module generates subset prototypes locating the feature space region of uncertain pixels. Soft multi-labels, computed using both subset and single-class prototypes, determine the relative position of uncertain pixels' feature vectors in the feature space. Leveraging soft multi-labels, our developed hard uncertain pixel mining (HM) module distinguishes easy uncertain samples, accepting their refined labels as trusted pseudo-labels. For hard uncertain samples, despite unreliable refined labels, some certainty to non-associated classes is observed. Thus, their soft multi-labels are stored in a memory bank as negative samples for certain classes. Then, contrastive learning is applied to all data to attract the positive sample pairs and repulse the negative sample pairs. We evaluated the proposed SCS-SLSP against several state-of-the-art semi-supervised approaches on three public datasets, and the results demonstrate the superiority of our network. When trained with 10% labeled data on the ACDC dataset, compared to the best results in all compared methods, SCS-SLSP improved the Dice by 0.41%, increased the JSC by 1.02%, reduced the HD95 by 56.82%, and decreased the ASD by 49.63%. The code for this work is available at https://github.com/helathegreat/SCS-SLSP. • Semi-supervised cardiac image segmentation using subjective logic theory and subset prototype generation. • Subjective logic theory avoids threshold setting, which can dynamically select certain/uncertain pixels. • SPG is designed to generate subset prototypes to specify the feature space region. • A novel soft multi-label fusion and hard uncertain pixel mining schemes to refine the easy uncertain pixels. [ABSTRACT FROM AUTHOR]

Copyright of Informatica (03505596) is the property of Slovene Society Informatika and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

Researchers and practitioners in various fields, including e-commerce customization, information retrieval, ubiquitous and mobile computing, data mining, marketing, and management, have realized the value of contextual information. Context-aware recommender systems assist users in finding their chosen material in a reasonable amount of time by utilizing information that describes the scenario in which the items will be consumed. For better personalized user recommendation, recommender systems leverage the contextual information in their process of recommendation called context-aware recommendation. Classification is used for context-prediction which represents the prediction of future context based on recorded previous context. The context prediction algorithm's goal is to recognize typical behavior patterns that have been seen in the past and then offer the most likely continuation of a presently observed collection of context components based on this knowledge. In this article we study the correlation between the multi-class classification and the context-aware recommendation.With this correlation we conclude that the linkage between contextual information and classification enhance and improve the recommendation results. [ABSTRACT FROM AUTHOR]

In the engineering practice, lacking of data especially labeled data typically hinders the wide application of deep learning in mechanical fault diagnosis. However, collecting and labeling data is often expensive and time-consuming. To address this problem, a kind of semi-supervised meta-learning networks (SSMN) with squeeze-and-excitation attention is proposed for few-shot fault diagnosis in this paper. SSMN consists of a parameterized encoder, a non-parameterized prototype refinement process and a distance function. Based on attention mechanism, the encoder is able to extract distinct features to generate prototypes and enhance the identification accuracy. With semi-supervised few-shot learning, SSMN utilizes unlabeled data to refine original prototypes for better fault recognition. A combinatorial learning optimizer is designed to optimize SSMN efficiently. The effectiveness of the proposed method is demonstrated through three bearing vibration datasets and the results indicate the outstanding adaptability in different situations. Comparison with other approaches is also made under the same setup and the experimental results prove the superiority of the proposed method for few-shot fault diagnosis. • Propose semi-supervised meta-learning networks (SSMN) with squeeze-and-excitation attention for few-shot fault diagnosis. • Generate discriminative features from raw inputs via attention mechanism. • Refine prototypes for better fault identification by semi-supervised meta-learning. • Combine SGD and Adam to converge the training loss efficiently for SSMN. • Conduct extensive experiments to evaluate the effectiveness by bearing vibration datasets. [ABSTRACT FROM AUTHOR]

Aiming at the problem of software defect prediction difficulty caused by insufficient software defect marker samples and unbalanced classification, a semi-supervised software defect prediction model based on a tri-training algorithm was proposed by combining feature normalization, over-sampling technology, and a Tri-training algorithm. First, the feature normalization method is used to smooth the feature data to eliminate the influence of too large or too small feature values on the model's classification performance. Secondly, the oversampling method is used to expand and sample the data, which solves the unbalanced classification of labelled samples. Finally, the Tri-training algorithm performs machine learning on the training samples and establishes a defect prediction model. The novelty of this model is that it can effectively combine feature normalization, oversampling techniques, and the Tri-training algorithm to solve both the under-labelled sample and class imbalance problems. Simulation experiments using the NASA software defect prediction dataset show that the proposed method outperforms four existing supervised and semi-supervised learning in terms of Precision, Recall, and F-Measure values. [ABSTRACT FROM AUTHOR]

The pulmonary nodules' malignancy rating is commonly confined in patient follow-up; examining the nodule's activity is estimated with the Positron Emission Tomography (PET) system or biopsy. However, these strategies are usually after the initial detection of the malignant nodules acquired from the Computed Tomography (CT) scan. In this study, a Deep Learning methodology to address the challenge of the automatic characterisation of Solitary Pulmonary Nodules (SPN) detected in CT scans is proposed. The research methodology is based on Convolutional Neural Networks, which have proven to be excellent automatic feature extractors for medical images. The publicly available CT dataset, called Lung Image Database Consortium and Image Database Resource Initiative (LIDC-IDRI), and a small CT scan dataset derived from a PET/CT system, is considered the classification target. New, realistic nodule representations are generated employing Deep Convolutional Generative Adversarial Networks to circumvent the shortage of large-scale data to train robust CNNs. Besides, a hierarchical CNN called Feature Fusion VGG19 (FF-VGG19) was developed to enhance feature extraction of the CNN proposed by the Visual Geometry Group (VGG). Moreover, the generated nodule images are separated into two classes by utilising a semi-supervised approach, called self-training, to tackle weak labelling due to DC-GAN inefficiencies. The DC-GAN can generate realistic SPNs, as the experts could only distinguish 23 % of the synthetic nodule images. As a result, the classification accuracy of FF-VGG19 on the LIDC-IDRI dataset increases by +7%, reaching 92.07 %, while the classification accuracy on the CT dataset is increased by 5 %, reaching 84,3 %. [ABSTRACT FROM AUTHOR]

In recent years, facial attribute editing has been successfully used to effectively change face images of various attributes based on generative adversarial networks and encoder-decoder models. However, existing models have a limitation in that they may change an unintended part in the process of changing an attribute or may generate an unnatural result. In this paper, we propose a model that improves the learning of the attention mask by adding a spatial attention mechanism based on the unified selective transfer network (referred to as STGAN) using semi-supervised learning. The proposed model can edit multiple attributes while preserving details independent of the attributes being edited. This study makes two main contributions to the literature. First, we propose an encoder-decoder model structure that learns and edits multiple facial attributes and suppresses distortion using an attention mask. Second, we define guide masks and propose a method and an objective function that use the guide masks for multiple facial attribute editing through semi-supervised learning. Through qualitative and quantitative evaluations of the experimental results, the proposed method was proven to yield improved results that preserve the image details by suppressing unintended changes than existing methods. [ABSTRACT FROM AUTHOR]

• A semi-supervised learning method for tooth segmentation on a dental CBCT dataset. • Use of a new inter-feature module - the Adaptive Channel Interaction Module (ACIM) in place of the skip connection. • A novel uncertainty loss function which enhances the constraint on the distance to the boundary. Automatic and accurate segmentation of teeth from CBCT is of great importance for dental diagnosis and treatment. In this paper, we propose a symmetric consistent semi-supervised segmentation network by uncertainty regularization that enables accurate segmentation of teeth with limited labeled data. Specifically, a two-stage segmentation framework is developed, where the input CBCT images are first passed through a coarse simple network to obtain the bounding box of the crown region, which is then fed into the proposed network to obtain a refined segmentation probability map. In particular, in place of the commonly used residual skip connection, we use Adaptive Channel Interaction Module (ACIM) between the encoder and decoder. Incremental feature fusion operations are performed in a left-to-right and bottom-to-top order to enhance feature reusing. To effectively utilize the unlabeled data, we combine the consistency between the pixel-by-pixel segmentation task and the level-set regression task. We use the average feature map as a new cue to compute the uncertainty of each layer as a regularization constraint term for predicting the segmentation. Experiments on the CBCT dental 3D models provided by Fudan dual-modality dental imaging (FDDI) dataset show that the proposed method outperforms existing semi-supervised learning methods in all five metrics considered, which demonstrates the effectiveness of the novel pipeline proposed in this work for the dental segmentation task. [ABSTRACT FROM AUTHOR]