Semi-supervised Learning Algorithm Research Articles

A semi-supervised learning approach with consistency regularization for tumor histopathological images analysis.

Manual inspection of histopathological images is important in clinical cancer diagnosis. Pathologists implement pathological diagnosis and prognostic evaluation through the microscopic examination of histopathological slices. This entire process is time-consuming, laborious, and challenging for pathologists. The modern use of whole-slide imaging, which scans histopathology slides to digital slices, and analysis using computer-aided diagnosis is an essential problem. To solve the problem of difficult labeling of histopathological data, and improve the flexibility of histopathological analysis in clinical applications, we herein propose a semi-supervised learning algorithm coupled with consistency regularization strategy, called"Semi- supervised Histopathology Analysis Network"(Semi-His-Net), for automated normal-versus-tumor and subtype classifications. Specifically, when inputted disturbing versions of the same image, the model should predict similar outputs. Based on this, the model itself can assign artificial labels to unlabeled data for subsequent model training, thereby effectively reducing the labeled data required for training. Our Semi-His-Net is able to classify patches from breast cancer histopathological images into normal tissue and three other different tumor subtypes, achieving an accuracy was 90%. The average AUC of cross-classification between tumors reached 0.893. To overcome the limitations of visual inspection by pathologists for histopathology images, such as long time and low repeatability, we have developed a deep learning-based framework (Semi-His-Net) for automatic classification subdivision of the subtypes contained in the whole pathological images. This learning-based framework has great potential to improve the efficiency and repeatability of histopathological image diagnosis.

Fast semi-supervised self-training algorithm based on data editing

Self-training is a commonly semi-supervised learning Algorithm framework. How to select the high-confidence samples is a crucial step for algorithms based on self-training framework. To alleviate the impact of noise data, researchers have proposed many data editing methods to improve the selection quality of high-confidence samples. However, the state-of-the-art data editing methods have high time complexity, which is not less than O(n2), where n denotes the number of samples. To improve the training speed while ensuring the quality of the selected high-confidence samples, inspired by Ball-k-means algorithm, we propose a fast semi-supervised self-training Algorithm based on data editing (EBSA), which defines ball-cluster partition and editing to improve the quality of high-confidence samples. The time complexity of the proposed EBSA is O(t2kn+nlogn+n+k2), where k denotes the number of centers, t denotes the number of iterates. k is far less than n, EBSA has linear time complexity with respect to n. A large number of experiments on 20 benchmark data sets have been carried out and the experimental results show that the proposed Algorithm not only ran faster, but also obtained better classification performance compared with the comparison algorithms.

Semi-supervised attribute reduction for partially labeled categorical data based on predicted label

In many practical applications of machine learning, there are a large number of partially labeled categorical data due to the high cost of labeling data. The semi-supervised learning algorithm is needed to deal with such data. This paper studies the label prediction of partially labeled categorical data and considers semi-supervised attribute reduction in a partially labeled categorical decision information system (p-CDIS) with predicted labels. The labels of unlabeled data are first predicted by means of the conditional probability. Then, uncertainty measurement for a p-CDIS with predicted labels is studied, and the dependence and conditional information entropy (CIE) are defined. Next, based on the dependence and CIE, two attribute reduction algorithms are designed. In addition, the effect of label deletion rate (LDR) on the dependence, CIE and reduction results are also studied. Finally, the results of experiments and statistical tests on 16 categorical UCI datasets show that the designed algorithms are statistically better than some state-of-the-art algorithms in classification accuracy.

Image classification of hyperspectral remote sensing using semi-supervised learning algorithm.

Hyperspectral images contain abundant spectral and spatial information of the surface of the earth, but there are more difficulties in processing, analyzing, and sample-labeling these hyperspectral images. In this paper, local binary pattern (LBP), sparse representation and mixed logistic regression model are introduced to propose a sample labeling method based on neighborhood information and priority classifier discrimination. A new hyperspectral remote sensing image classification method based on texture features and semi-supervised learning is implemented. The LBP is employed to extract features of spatial texture information from remote sensing images and enrich the feature information of samples. The multivariate logistic regression model is used to select the unlabeled samples with the largest amount of information, and the unlabeled samples with neighborhood information and priority classifier discrimination are selected to obtain the pseudo-labeled samples after learning. By making full use of the advantages of sparse representation and mixed logistic regression model, a new classification method based on semi-supervised learning is proposed to effectively achieve accurate classification of hyperspectral images. The data of Indian Pines, Salinas scene and Pavia University are selected to verify the validity of the proposed method. The experiment results have demonstrated that the proposed classification method is able to gain a higher classification accuracy, a stronger timeliness, and the generalization ability.

A Reliable Application of MPC for Securing the Tri-Training Algorithm

Due to the widespread use of distributed data mining techniques in a variety of areas, the issue of protecting the privacy of sensitive data has received increasing attention in recent years. Privacy-preserving distributed data mining (PPDDM) focuses on decentralized data analysis without the disclosure of sensitive information from data owner. However, the previous PPDDM mostly works on a limited amount of labeled data. In contrast to the real world, unlabeled data is abundance and labeled data is scarce. The objectives of this paper are to study and to analyze privacy-preserving properties of semi-supervised learning (SSL) algorithm with the combination of labeled and unlabeled data, where data is distributed among multiple data owners. In this paper we propose a Privacy-preserving Distributed Data Mining (PPDDM) method by designing a reliable application of secure MPC to semi-supervised tri-training algorithms. We simulate the original tri-training algorithm and tri-training algorithm with secure MPC using a different types of classifiers and datasets. The simulation results show that tri-training in secure MPC has almost same accuracy compared to original tri-training algorithm. We also compare execution time in addition to performance evaluation of tri-training in secure and the original tri-training algorithms.

A unified network embedding algorithm for multi-type similarity measures

Traditional network embedding aims to learn representations by capturing a predefined vertex-to-vertex similarity measure. However, in practice, there are different types of similarity measures (e.g., connectivity and structural similarity), which are appropriate for different downstream applications. Meanwhile, it is hard to select the “best” similarity measure that can mostly benefit the application, considering the required domain knowledge of both application scenario and network science. It sometimes requires to cooperate these similarity measures with each other for achieving better performance. Therefore, automatically integrate multiple types of similarity measures into a uniform network embedding framework is critical to obtain effective vertex representations for a downstream application. In this paper, we address the above problem in social networks, and propose a semi-supervised representation learning algorithm. The general idea of our approach is to impose social influence, which occurs when one’s opinions, emotions, or behaviors are affected by others in a social network. Particularly, we build the connection between a user’s representation vector and the probability of her being influenced by another user to have a particular label (e.g., fraud, personal interest, etc.). We conduct efficient experiments based on six real-world datasets and find a clear improvement of our approach comparing with several state-of-the-art baselines.

Analysis of Machine-Based Learning Algorithm Used in Named Entity Recognition

Aim/Purpose: The amount of information published has increased dramatically due to the information explosion. The issue of managing information as it expands at this rate lies in the development of information extraction technology that can turn unstructured data into organized data that is understandable and controllable by computers Background: The primary goal of named entity recognition (NER) is to extract named entities from amorphous materials and place them in pre-defined semantic classes. Methodology: In our work, we analyze various machine learning algorithms and implement K-NN which has been widely used in machine learning and remains one of the most popular methods to classify data. Contribution: To the researchers’ best knowledge, no published study has presented Named entity recognition for the Kikuyu language using a machine learning algorithm. This research will fill this gap by recognizing entities in the Kikuyu language. Findings: An evaluation was done by testing precision, recall, and F-measure. The experiment results demonstrate that using K-NN is effective in classification performance. Recommendation for Researchers: With enough training data, researchers could perform an experiment and check the learning curve with accuracy that compares to state of art NER. Future Research: Future studies may be done using unsupervised and semi-supervised learning algorithms for other resource-scarce languages.

Methodology of semi-supervised algorithm selection for classification problems

The paper concerns the problem of selecting an appropriate semi-supervised learning algorithm based on validating assumptions that the algorithm is based on for the particular dataset. This enables the selection of the fittest algorithm to achieve the best possible model accuracy. In this paper, we provide an overview of four primary semi-supervised assumptions, introduce the definition of metrics used to evaluate the assumption, provide a taxonomy of common semi-supervised algorithms and assumptions based on, and evaluate the proposed methodology on the synthetic too-moons dataset. The results indicate prospects of improving methodologies further by refining and introducing new metrics.

Comparative analysis of modified semi-supervised learning algorithms on a small amount of labeled data

The paper is devoted to improving semi-supervised clustering methods and comparing their accuracy and robustness. The proposed approach is based on expanding a clustering algorithm for using an available set of labels by replacing the distance function. Using the distance function considers not only spatial data but also available labels. Moreover, the proposed distance function could be adopted for working with ordinal variables as labels. An extended approach is also considered, based on a combination of unsupervised k-medoids methods, modified for using only labeled data during the medoids calculation step, supervised method of k nearest neighbor, and unsupervised k-means. The learning algorithm uses information about the nearest points and classes’ centers of mass. The results demonstrate that even a small amount of labeled data allows us to use semi-supervised learning, and proposed modifications improve accuracy and algorithm performance, which was found during experiments.

Retraction Note: A novel semi supervised learning algorithm for thyroid and ulcer classification in tongue image

Retraction Note: A novel semi supervised learning algorithm for thyroid and ulcer classification in tongue image

Data Classification and Demand Prediction Methods Based on Semi-Supervised Agricultural Machinery Spare Parts Data

Because of the continuous improvement of technology, mechanization has emerged in various fields. Due to the different suitable seasons for the growth of agricultural plants, agricultural mechanization faces problems different from other industries. That is, agricultural machinery and equipment may be used frequently for a period of time, or may be idle for a long time. This leads to the aging of equipment no longer becoming regular, the maintenance time of spare parts is not fixed, the number of spare parts stored in the spare parts warehouse cannot be too large to occupy funds, and the number cannot be too small to meet the maintenance needs, so the prediction of agricultural machinery spare parts has become particularly important. Due to the lack of information, the difficulty of labeling, and the imbalance of positive and negative sample classification, this paper used a semi-supervised learning algorithm to solve the problem of agricultural machinery spare parts data classification. In order to forecast the demand for spare parts of agricultural machinery, this paper compared the IPSO-BP neural network algorithm and BP neural network algorithm. It was found that the IPSO-BP neural network was used to forecast the demand for spare parts of agricultural machinery, and the error between the predicted value and the actual value was small and met the accuracy requirements.

The High Separation Probability Assumption for Semi-Supervised Learning

We propose a novel semi-supervised learning (SSL) assumption—the high separation probability (HSP) assumption—that is complementary to the common low density separation (LDS) assumption, and a highly accurate and efficient SSL algorithm, the transductive minimax probability machine (TMPM). Under the HSP assumption, a preferred classifier should ensure that with high probability the samples are far away from the decision boundary. Similarly, the LDS assumption states that the decision boundary should lie in a low-density region. A remarkable property of HSP is that it provides an elegant theoretical framework for analyzing the generalization performance of SSL, while it remains difficult for the LDS assumption to have a quantitative criterion on how it fits the data. Moreover, it has the same paradigm as the minimax probability machine (MPM) and yields an intractable mixed-integer optimization for solving the transductive label assignment problem. Consequently, we propose TMPM with a label-switching strategy and iterative manner to formalize our approximate solution. Theoretically, we analyze the computational complexity, convergence property, and generalization error bound of TMPM. Empirically, we show that TMPM achieves competitive performance on a wide range of datasets, while being almost one order of magnitude faster than existing SSL approaches.

Semi‐supervised learning dehazing algorithm based on the OSV model

Despite the great progress that has been made in the task of single image dehazing, the results of the existing models in restoring image edge and texture information are still challenging. Besides, most dehazing models are trained on synthetic data, resulting in poor generalization ability to real-world images. To address the aforementioned problems, a semi-supervised learning dehazing method based on the decomposition model of Osher, Solé, and Vese(The OSV model) is presented. Specifically, the OSV model is first applied to decompose the hazy image into the structure layer and texture layer, save the texture layer and dehaze for the structure layer to restore images with sharper texture and edge. Furthermore, the network adopts a semi-supervised learning algorithm based on generative adversarial networks (GAN) to generalize better to real-world images, which includes two branches: supervised learning and unsupervised learning. Extensive experiments indicate that the proposed method preserves the texture and edge information of images more accurately while dehazing better, and performs favourably against the advanced dehazing algorithms on both synthetic outdoor datasets and real-world hazy images.

Semisupervised Radar-Based Gait Recognition in the Wild via Ensemble Determination Strategy

Radar-based gait recognition has attracted growing interest due to radar’s benefit of robust to harsh environments, and active researchers have recently studied the feasibility of semisupervised radar-based tasks to reduce the cost of labeling. However, in order to obtain sufficient radar data to meet the requirements of semisupervised learning and to obtain good generalized gait recognition models, radar data need to be collected for various cases, resulting in the “independently and identically distributed (i.i.d.) assumption” in semisupervised learning that cannot be satisfied due to the presence of “covariant perturbations.” In this article, we propose a semisupervised model for radar-based gait recognition in the wild, and this model mitigates the performance degradation observed in the non-i.i.d. semisupervised condition. The model assesses the extent to which unlabeled samples are influenced by “covariant perturbations” and then differentiates how they are incorporated in semisupervised training, which not only prevents the samples whose semantics are destroyed from training but also enables the model to place more emphasis on the “hard samples.” We create a radar measurement dataset for the performance evaluation, and our model surpasses other seven semisupervised learning algorithms, demonstrating the effectiveness of our model in radar-based gait recognition problems.

Adaptive Weighted Losses With Distribution Approximation for Efficient Consistency-Based Semi-Supervised Learning

Recent semi-supervised learning (SSL) algorithms such as FixMatch achieve state-of-the-art performance by exploiting consistency regularization and entropy minimization techniques. However, many consistency-based SSL algorithms extract pseudo-labels from unlabeled data through a fixed threshold and ignore the different learning progress of each category, which makes the easy-to-learn categories have more examples contributing to the loss, resulting in a class-imbalance problem and affecting training efficiency. In order to improve the training reliability, we propose adaptive weighted losses (AWL). Through the evaluation of the class-wise learning progress, the loss contribution of the pseudo-labeled data of each category is continuously and dynamically adjusted during the learning process, and the pseudo-label discrimination ability of the model can be steadily improved. Moreover, to improve the training efficiency, we propose a bidirectional distribution approximation (DA) method, which introduces the consistency information of the predictions under the threshold into the loss calculation, and significantly improves the model convergence speed. Through the combination of AWL and DA, our method surpasses the performance of other algorithms on multiple benchmarks with a faster convergence efficiency, especially in the case of labeled data extremely limited. For example, AWL&DA achieves 95.29% test accuracy on the CIFAR-10-40-labels experiment and 92.56% accuracy on a faster experiment setting with only <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$2^{18}$ </tex-math></inline-formula> iterations.

Semi-supervised health assessment of pipeline systems based on optical fiber monitoring

Health assessment of pipeline systems is of deep significance for improving pipeline reliability and integrity. Traditional health assessment methods may be difficult or costly to perform on pipeline systems due to the long distances and environmental constraints of pipelines. This paper incorporates the distributed optical fiber sensor (DOFS) technique and the semi-supervised learning algorithm into the pipeline health assessment framework. Three critical problems that limit the application of DOFS in pipeline health assessments are addressed. First, an applicable damage monitoring experiment of a pipeline system is designed, which is effective in obtaining the necessary base data for data-driven modeling. Second, the correspondence between the pipeline health status and the monitored strain features is established. The experimental data are shared for public research, which is expected to solve the problem of the lack of benchmark research data in related fields. Third, considering the scarcity of labeled degradation data in pipelines, a semi-supervised denoising autoencoder model is proposed specifically for pipeline health assessment. The proposed method is demonstrated and validated using a comparative experimental case study.

A semi-supervised adaptive Markov Gaussian embedding process (SAMGEP) for prediction of phenotype event times using the electronic health record

While there exist numerous methods to identify binary phenotypes (i.e. COPD) using electronic health record (EHR) data, few exist to ascertain the timings of phenotype events (i.e. COPD onset or exacerbations). Estimating event times could enable more powerful use of EHR data for longitudinal risk modeling, including survival analysis. Here we introduce Semi-supervised Adaptive Markov Gaussian Embedding Process (SAMGEP), a semi-supervised machine learning algorithm to estimate phenotype event times using EHR data with limited observed labels, which require resource-intensive chart review to obtain. SAMGEP models latent phenotype states as a binary Markov process, and it employs an adaptive weighting strategy to map timestamped EHR features to an embedding function that it models as a state-dependent Gaussian process. SAMGEP’s feature weighting achieves meaningful feature selection, and its predictions significantly improve AUCs and F1 scores over existing approaches in diverse simulations and real-world settings. It is particularly adept at predicting cumulative risk and event counting process functions, and is robust to diverse generative model parameters. Moreover, it achieves high accuracy with few (50–100) labels, efficiently leveraging unlabeled EHR data to maximize information gain from costly-to-obtain event time labels. SAMGEP can be used to estimate accurate phenotype state functions for risk modeling research.

Semi-supervised partial label learning algorithm via reliable label propagation

Semi-supervised partial label learning algorithm via reliable label propagation

FCCCSR_Glu: a semi-supervised learning model based on FCCCSR algorithm for prediction of glutarylation sites.

Glutarylation is a post-translational modification which plays an irreplaceable role in various functions of the cell. Therefore, it is very important to accurately identify the glutarylation substrates and its corresponding glutarylation sites. In recent years, many computational methods of glutarylation sites have emerged one after another, but there are still many limitations, among which noisy data and the class imbalance problem caused by the uncertainty of non-glutarylation sites are great challenges. In this study, we propose a new semi-supervised learning algorithm, named FCCCSR, to identify reliable non-glutarylation lysine sites from unlabeled samples as negative samples. FCCCSR first finds core objects from positive samples according to reverse nearest neighbor information, and then clusters core objects based on natural neighbor structure. Finally, reliable negative samples are selected according to clustering result. With FCCCSR algorithm, we propose a new method named FCCCSR_Glu for glutarylation sites identification. In this study, multi-view features are extracted and fused to describe peptides, including amino acid composition, BLOSUM62, amino acid factors and composition of k-spaced amino acid pairs. Then, reliable negative samples selected by FCCCSR and positive samples are combined to establish models and XGBoost optimized by differential evolution algorithm is used as the classifier. On the independent testing dataset, FCCCSR_Glu achieves 85.18%, 98.36%, 94.31% and 0.8651 in sensitivity, specificity, accuracy and Matthew's Correlation Coefficient, respectively, which is superior to state-of-the-art methods in predicting glutarylation sites. Therefore, FCCCSR_Glu can be a useful tool for glutarylation sites prediction and FCCCSR algorithm can effectively select reliable negative samples from unlabeled samples. The data and code are available on https://github.com/xbbxhbc/FCCCSR_Glu.git.

An Improved Algorithm of Drift Compensation for Olfactory Sensors

This research mainly studies the semi-supervised learning algorithm of different domain data in machine olfaction, also known as sensor drift compensation algorithm. Usually for this kind of problem, it is difficult to obtain better recognition results by directly using the semi-supervised learning algorithm. For this reason, we propose a domain transformation semi-supervised weighted kernel extreme learning machine (DTSWKELM) algorithm, which converts the data through the domain and uses SWKELM algorithmic classification to transform the semi-supervised classification problem of different domain data into a semi-supervised classification problem of the same domain data.