Unlabeled Database Research Articles

Text Classification of Patient Experience Comments in Saudi Dialect Using Deep Learning Techniques

Improving the quality of healthcare services is of the utmost importance in healthcare systems. Patient experience is a key aspect that should be gauged and monitored continuously. However, the measurement of such a vital indicator typically cannot be carried out directly, instead being derived from the opinions of patients who usually express their experience in free text. When it comes to patient comments written in the Arabic language, the currently used strategy to classify Arabic comments is totally reliant on human annotation, which is time-consuming and prone to subjectivity and error. Thus, fully using the value of patient feedback in a timely manner is difficult. This paper addresses the problem of classifying patient experience (PX) comments written in Arabic into 25 classes by using deep learning- and BERT-based models. A real-world data set of patient comments is obtained from the Saudi Ministry of Health for this purpose. Features are extracted from the data set, then used to train deep learning-based classifiers—including BiLSTM and BiGRU—for which pre-trained static word embedding and pre-training vector word embeddings are utilized. Furthermore, we utilize several Arabic pre-trained BERT models, in addition to building PX_BERT, a customized BERT model using the PX unlabeled database. From the experimental results for the 28 classifiers built in this study, the best-performing models (based on the F1 score) are found to be PX_BERT and AraBERTv02. To the best of our knowledge, this is the first study to tackle PX comment classification for the Arabic language.

Enforcing Semantic Consistency for Cross Corpus Emotion Prediction Using Adversarial Discrepancy Learning in Emotion

Mismatch between databases entails a challenge in performing emotion recognition on a practical-condition unlabeled database with labeled source data. The alignment between the source and target is crucial for conventional neural network; therefore, many studies have mapped two domains in a common feature space. However, the effect of distortion in emotion semantics across different conditions has been neglected in such work, and a sample from the target may be considered a high emotional annotation in the target but as low in the source. In this work, we propose the maximum regression discrepancy (MRD) network, which enforces semantic consistency in a source and target by adjusting the acoustic feature encoder to minimize discrepancy in maximally distorted samples through adversarial training. We show our framework in several experiments using three databases (the USC IEMOCAP, MSP-Improv, and MSP-Podcast) for cross corpus emotion prediction. Compared to the Source-only neural network and DANN, MRD network demonstrates a significant improvement between 5\% and 10\% in the concordance correlation coefficient (CCC) in cross-corpus prediction and between 3\% and 10\% for evaluation on MSP-PODCAST. We also visualize the effect of MRD on feature representation to shows the efficacy of the MRD structure we designed.

Learning Semantic Features for Classifying Very Large Image Datasets Using Convolution Neural Network

Advancements in sensors and image acquisition devices lead to tremendous increase in creation of unlabeled database of images, and traditional image retrieval approaches are inefficient in retrieving semantic images. Neural network is emerging as popular method to solve most of the state-of-the-art problems in filling the gap between low level features and high level semantics. Convolution neural network (CNN) is a category of neural network which automatically extracts the important features without any human intervention, with considerably reduced set of parameters. In this paper, a model is generated using CNN (VGG-16) architecture which combines convolution and max pooling layers at different levels using effective regularization and transfer learning with data augmentation. The effectiveness of proposed model is demonstrated on benchmarking Caltech 101 and 256 datasets. The results obtained have proven the capacity of model in understanding high level semantics and outperformed several contemporary techniques.

Detection of Alfvén Eigenmodes on COMPASS with Generative Neural Networks

Chirping Alfvén eigenmodes were observed at the COMPASS tokamak. They are believed to be driven by runaway electrons (REs), and as such, they provide a unique opportunity to study the physics of nonlinear interaction between REs and electromagnetic instabilities, including important topics of RE mitigation and losses. On COMPASS, they can be detected from spectrograms of certain magnetic probes. So far, their detection has required much manual effort since they occur rarely. We strive to automate this process using machine learning techniques based on generative neural networks. We present two different models that are trained using a smaller, manually labeled database and a larger unlabeled database from COMPASS experiments. In a number of experiments, we demonstrate that our approach is a viable option for automated detection of rare instabilities in tokamak plasma.

Unsupervised learning for large-scale corneal topography clustering

Machine learning algorithms have recently shown their precision and potential in many different use cases and fields of medicine. Most of the algorithms used are supervised and need a large quantity of labeled data to achieve high accuracy. Also, most applications of machine learning in medicine are attempts to mimic or exceed human diagnostic capabilities but little work has been done to show the power of these algorithms to help collect and pre-process a large amount of data. In this study we show how unsupervised learning can extract and sort usable data from large unlabeled datasets with minimal human intervention. Our digital examination tools used in clinical practice store such databases and are largely under-exploited. We applied unsupervised algorithms to corneal topography examinations which remains the gold standard test for diagnosis and follow-up of many corneal diseases and refractive surgery screening. We could extract 7019 usable examinations which were automatically sorted in 3 common diagnoses (Normal, Keratoconus and History of Refractive Surgery) from an unlabeled database with an overall accuracy of 96.5%. Similar methods could be used on any form of digital examination database and greatly speed up the data collection process and yield to the elaboration of stronger supervised models.

Towards automatic setup of non intrusive appliance load monitoring – feature extraction and clustering

<p style="-qt-block-indent: 0; text-indent: 0px; margin: 0px;">Given climate change concerns and incessantly increasing energy demands of the present time, improving energy efficiency becomes of significant environmental and economic impact. Monitoring household electrical consumption through a non-intrusive appliance load monitoring (NIALM) system achieves significant efficiency improvement by providing appliance-level energy consumption and relaying this information back to the user. This paper focuses on feature extraction and clustering, which constitute two of the four modules of the proposed automatic-setup NIALM system, the other two being labeling and classification. The feature extraction module applies the Estimation of Signal Parameters via Rotational Invariance Techniques (ESPRIT), a well-known parametric estimation technique, to the drawn electric current. The result is a compact representation of the signal in terms of complex numbers referred to as poles and residues. These complex numbers are then used to determine a feature vector consisting of the contribution of the fundamental, the third and the fifth harmonic currents to the maximum of the total load current. Once a signature is extracted, the clustering module applies distance-based rules inferred off-line from various databases and decides either to create a new class out of the new signature or to discard it and increase the count of an existing signature. As a result, the feature space is clustered without the a priori knowledge of the number of appliances into singleton clusters. Results obtained from a set of appliances indicate that these two modules succeed in creating an unlabeled database of signatures.</p>

Retinal Image Synthesis and Semi-Supervised Learning for Glaucoma Assessment.

Recent works show that generative adversarial networks (GANs) can be successfully applied to image synthesis and semi-supervised learning, where, given a small labeled database and a large unlabeled database, the goal is to train a powerful classifier. In this paper, we trained a retinal image synthesizer and a semi-supervised learning method for automatic glaucoma assessment using an adversarial model on a small glaucoma-labeled database and a large unlabeled database. Various studies have shown that glaucoma can be monitored by analyzing the optic disc and its surroundings, and for that reason, the images used in this paper were automatically cropped around the optic disc. The novelty of this paper is to propose a new retinal image synthesizer and a semi-supervised learning method for glaucoma assessment based on the deep convolutional GANs. In addition, and to the best of our knowledge, this system is trained on an unprecedented number of publicly available images (86926 images). This system, hence, is not only able to generate images synthetically but to provide labels automatically. Synthetic images were qualitatively evaluated using t-SNE plots of features associated with the images and their anatomical consistency was estimated by measuring the proportion of pixels corresponding to the anatomical structures around the optic disc. The resulting image synthesizer is able to generate realistic (cropped) retinal images, and subsequently, the glaucoma classifier is able to classify them into glaucomatous and normal with high accuracy (AUC = 0.9017). The obtained retinal image synthesizer and the glaucoma classifier could then be used to generate an unlimited number of cropped retinal images with glaucoma labels.

Multi-feature classifiers for burst detection in single EEG channels from preterm infants

Objective. The study of electroencephalographic (EEG) bursts in preterm infants provides valuable information about maturation or prognostication after perinatal asphyxia. Over the last two decades, a number of works proposed algorithms to automatically detect EEG bursts in preterm infants, but they were designed for populations under 35 weeks of post menstrual age (PMA). However, as the brain activity evolves rapidly during postnatal life, these solutions might be under-performing with increasing PMA. In this work we focused on preterm infants reaching term ages (PMA ⩾36 weeks) using multi-feature classification on a single EEG channel. Approach. Five EEG burst detectors relying on different machine learning approaches were compared: logistic regression (LR), linear discriminant analysis (LDA), k-nearest neighbors (kNN), support vector machines (SVM) and thresholding (Th). Classifiers were trained by visually labeled EEG recordings from 14 very preterm infants (born after 28 weeks of gestation) with 36–41 weeks PMA. Main results. The most performing classifiers reached about 95% accuracy (kNN, SVM and LR) whereas Th obtained 84%. Compared to human-automatic agreements, LR provided the highest scores (Cohen’s kappa = 0.71) using only three EEG features. Applying this classifier in an unlabeled database of 21 infants ⩾36 weeks PMA, we found that long EEG bursts and short inter-burst periods are characteristic of infants with the highest PMA and weights. Significance. In view of these results, LR-based burst detection could be a suitable tool to study maturation in monitoring or portable devices using a single EEG channel.

Universum Autoencoder-Based Domain Adaptation for Speech Emotion Recognition

One of the serious obstacles to the applications of speech emotion recognition systems in real-life settings is the lack of generalization of the emotion classifiers. Many recognition systems often present a dramatic drop in performance when tested on speech data obtained from different speakers, acoustic environments, linguistic content, and domain conditions. In this letter, we propose a novel unsupervised domain adaptation model, called Universum autoencoders, to improve the performance of the systems evaluated in mismatched training and test conditions. To address the mismatch, our proposed model not only learns discriminative information from labeled data, but also learns to incorporate the prior knowledge from unlabeled data into the learning. Experimental results on the labeled Geneva Whispered Emotion Corpus database plus other three unlabeled databases demonstrate the effectiveness of the proposed method when compared to other domain adaptation methods.

Image descriptor extraction using Android mobile devices for a video search system

The increasing amount of multimedia data makes it challenging to search for content, specially over an unlabeled database. Meanwhile, the proliferation of smartphones, which possess both video cameras and powerful processors, makes it possible to use these devices as part of a video search process. This paper presents a system for performing video searches using an Android mobile device, that records a video and extract one of three different kinds of global image descriptors, then it sends them to a server to perform a similarity search over a database. The precision of each kind of descriptor was measured. The efficiency of the system was compared with an analogous implementation which sends entire videos to the server and it was found to improve the system's response time and data usage.

Integrating multi-source Web records into relational database

How to integrate heterogeneous semi-structured Web records into relational database is an important and challengeable research topic. An improved model of conditional random fields was presented to combine the learning of labeled samples and unlabeled database records in order to reduce the dependence on tediously hand-labeled training data. The proposed model was used to solve the problem of schema matching between data source schema and database schema. Experimental results using a large number of Web pages from diverse domains show the novel approach's effectiveness.