Performance Of Semi-supervised Learning Research Articles

Semi-supervised feature selection analysis with structured multi-view sparse regularization

Facing abundant and various multi-view data, how to effectively combine the multi-view data information has become an important research topic in feature selection analysis. However, existing feature selection methods usually consider each view features as a whole without fully considering the individual feature in each view. In this paper, we construct a structured multi-view sparse regularization and then propose a novel semi-supervise feature selection framework, namely Structured Multi-view Hessian sparse Feature Selection (SMHFS)11SMHFS: Structured Multi-view Hessian sparse Feature Selection.. With the structured multi-view sparse regularization, SMHFS can simultaneously learn the importance of each view features and the importance of individual feature in each view. In addition, SMHFS utilizes multi-view Hessian regularization to enhance the semi-supervised learning performance. An iterative algorithm is introduced and its convergence is proven. Finally, SMHFS is applied into image annotation task and extensive experiments are conducted. The experimental results show SMHFS can effectively combine the multi-view data information to achieve better feature selection performance compared to other methods.

Safe Semi-Supervised Extreme Learning Machine for EEG Signal Classification

One major challenge in the current brain–computer interface research is the accurate classification of time-varying electroencephalographic (EEG) signals. The labeled EEG samples are usually scarce, while the unlabeled samples are available in large quantities and easy to collect in real applications. Semi-supervised learning (SSL) methods can utilize both labeled and unlabeled data to improve performance over supervised approaches. However, it has been reported that the unlabeled data may undermine the performance of SSL in some cases. To improve the safety of SSL, we proposed a new safety-control mechanism by analyzing the differences between unlabeled data analysis in supervised and semi-supervised learning. We then develop and implement a safe classification method based on the semi-supervised extreme learning machine (SS-ELM). Following this approach, the Wasserstein distance is used to measure the similarities between the predictions obtained from ELM and SS-ELM algorithms, and a different risk degree is thereby calculated for each unlabeled data instance. A risk-based regularization term is then constructed and embedded into the objective function of the SS-ELM. Extensive experiments were conducted using benchmark and EEG datasets to evaluate the effectiveness of the proposed method. Experimental results show that the performance of the new algorithm is comparable to SS-ELM and superior to ELM on average. It is thereby shown that the proposed method is safe and efficient for the classification of EEG signals.

Statistical-Mechanical Analysis Connecting Supervised Learning and Semi-Supervised Learning

The generalization performance of semi-supervised learning is analyzed in the framework of online learning using the statistical-mechanical method. We derive deterministically formed simultaneous differential equations that describe the dynamical behaviors of order parameters using the self-averaging property under the thermodynamic limit. By generalizing the number of labeled data, the derived theory connects supervised learning and semi-supervised learning.

Semisupervised Multiclass Classification Problems With Scarcity of Labeled Data: A Theoretical Study.

In recent years, the performance of semisupervised learning (SSL) has been theoretically investigated. However, most of this theoretical development has focused on binary classification problems. In this paper, we take it a step further by extending the work of Castelli and Cover to the multiclass paradigm. In particular, we consider the key problem in SSL of classifying an unseen instance x into one of K different classes, using a training data set sampled from a mixture density distribution and composed of l labeled records and u unlabeled examples. Even under the assumption of identifiability of the mixture and having infinite unlabeled examples, labeled records are needed to determine the K decision regions. Therefore, in this paper, we first investigate the minimum number of labeled examples needed to accomplish that task. Then, we propose an optimal multiclass learning algorithm, which is a generalization of the optimal procedure proposed in the literature for binary problems. Finally, we make use of this generalization to study the probability of error when the binary class constraint is relaxed.

Towards designing risk-based safe Laplacian Regularized Least Squares

Recently, Safe Semi-Supervised Learning (S3L) has become an active topic in the Semi-Supervised Learning (SSL) field. In S3L, unlabeled data that may affect the performance of SSL both positively and negatively are exploited more safely through different risk-based strategies, and such S3L methods are expected to perform at least the same as the corresponding Supervised Learning (SL) methods. While the previously proposed S3L methods considered the risk of unlabeled data, they did not explicitly model the different risk degrees of unlabeled data on the learning procedure. Hence, we propose risk-based safe Laplacian Regularized Least Squares (RsLapRLS) by analyzing the different risk degrees of unlabeled data in this paper. Our motivation is that unlabeled data may be risky in SSL and the risk degrees are different. We assign different risk degrees to unlabeled data according to the different characteristics in supervised and semi-supervised learning. Then a risk-based tradeoff term between supervised and semi-supervised learning is integrated into the objective function of SSL. The role of risk degrees is to determine the way of exploiting the unlabeled data. Unlabeled data with large risk degrees should be exploited by SL and others by SSL. In particular, we employ Regularized Least Squares (RLS) and Laplacian RLS (LapRLS) for SL and SSL, respectively. Experimental results on several UCI and benchmark datasets show that the performance of our algorithm is never significantly inferior to RLS and LapRLS. In this way, our algorithm improves the practicability of SSL.

Assessment of the equivalency for Troponin I as either negative or positive by Alere Triage and Abbott i-STAT relative to Siemens Dimension Vista

Recently, Safe Semi-Supervised Learning (S3L) has become an active topic in the Semi-Supervised Learning (SSL) field. In S3L, unlabeled data that may affect the performance of SSL both positively and negatively are exploited more safely through different risk-based strategies, and such S3L methods are expected to perform at least the same as the corresponding Supervised Learning (SL) methods. While the previously proposed S3L methods considered the risk of unlabeled data, they did not explicitly model the different risk degrees of unlabeled data on the learning procedure. Hence, we propose risk-based safe Laplacian Regularized Least Squares (RsLapRLS) by analyzing the different risk degrees of unlabeled data in this paper. Our motivation is that unlabeled data may be risky in SSL and the risk degrees are different. We assign different risk degrees to unlabeled data according to the different characteristics in supervised and semi-supervised learning. Then a risk-based tradeoff term between supervised and semi-supervised learning is integrated into the objective function of SSL. The role of risk degrees is to determine the way of exploiting the unlabeled data. Unlabeled data with large risk degrees should be exploited by SL and others by SSL. In particular, we employ Regularized Least Squares (RLS) and Laplacian RLS (LapRLS) for SL and SSL, respectively. Experimental results on several UCI and benchmark datasets show that the performance of our algorithm is never significantly inferior to RLS and LapRLS. In this way, our algorithm improves the practicability of SSL.

A risk degree-based safe semi-supervised learning algorithm

Semi-supervised learning has attracted much attention in machine learning field over the past decades and a number of algorithms are proposed to improve the performance by exploiting unlabeled data. However, unlabeled data may hurt performance of semi-supervised learning in some cases. It is instinctively expected to design a reasonable strategy to safety exploit unlabeled data. To address the problem, we introduce a safe semi-supervised learning by analyzing the different characteristics of unlabeled data in supervised and semi-supervised learning. Our intuition is that unlabeled data may be often risky in semi-supervised setting and the risk degree are different. Hence, we assign different risk degree to unlabeled data and the risk degree serve as a sieve to determine the exploiting way of unlabeled data. The unlabeled data with high risk should be exploited by supervised learning and the other should be used for semi-supervised learning. In particular, we utilize kernel minimum squared error (KMSE) and Laplacian regularized KMSE for supervised and semi-supervised learning, respectively. Experimental results on several benchmark datasets illustrate the performance of our algorithm is never inferior to that of KMSE and indicate the effectiveness and efficiency of our algorithm.