LLDSNet: Limited labeled data segmentation network for COVID-19 CT lesion segmentation

With the COVID-19 outbreak in 2019, the world is facing a major crisis and people's health is at serious risk. Accurate segmentation of lesions in CT images can help doctors understand disease infections, prescribe the right medicine and control patients' conditions. Fast and accurate diagnosis not only can make the limited medical resources get reasonable allocation, but also can control the spread of disease, and computer-aided diagnosis can achieve this purpose, so this paper proposes a deep learning segmentation network LLDSNet based on a small amount of data, which is divided into two modules: contextual feature-aware module (CFAM) and shape edge detection module (SEDM). Due to the different morphology of lesions in different CT, lesions with dispersion, small lesion area and background area imbalance, lesion area and normal area boundary blurred, etc. The problem of lesion segmentation in COVID-19 poses a major challenge. The CFAM can effectively extract the overall and local features, and the SEDM can accurately find the edges of the lesion area to segment the lesions in this area. The hybrid loss function is used to avoid the class imbalance problem and improve the overall network performance. It is demonstrated that LLDSNet dice achieves 0.696 for a small number of data sets, and the best performance compared to five currently popular segmentation networks. © 2023 SPIE.

FedUSC: Collaborative Unsupervised Representation Learning from Decentralized Data for Internet of Things

Federated Learning (FL) lately has shown much promise in improving the shared model and preserving data privacy. However, these existing methods are only of limited utility in the Internet of Things (IoT) scenarios, as they either heavily depend on high-quality labeled data or only perform well under idealized conditions, which typically cannot be found in practical applications. In this paper, we propose a novel federated unsupervised learning method for image classification without the use of any ground truth annotations. In IoT scenarios, a big challenge is that decentralized data among multiple clients is normally non-IID, leading to performance degradation. To address this issue, we further propose a dynamic update mechanism that can decide how to update the local model based on weights divergence. Extensive experiments show that our method outperforms all baseline methods by large margins, including +6.67% on CIFAR-10, +5.15% on STL-10, and +8.44% on SVHN in terms of classification accuracy. In particular, we obtain promising results on Mini-ImageNet and COVID-19 datasets and outperform several federated unsupervised learning methods under non-IID settings. IEEE

A COVID-19 detection method based on self-supervised learning and transfer learning

Since the global outbreak of COVID-19, the epidemic has had a great impact on people's lives and the world economy. Diagnosis of COVID-19 using deep learning has become increasingly important due to the inefficiency of traditional RT-PCR test. However, training deep neural networks requires a large amount of manually labeled data, and collecting a large number of COVID-19 CT images is difficult. To address this issue, we explore the effect of Pretext-Invariant Representation Learning (PIRL) using unlabeled datasets to pre-train the network on classification results. In addition, we also explore the prediction effect of PIRL combined with transfer learning (TF). According to the experimental results, applying the TF-PIRL prediction model constructed in this paper to COVID-19 diagnosis, the accuracy and AUC are 0.7734 and 0.8556 respectively, which outperform the network training from scratch, transfer learning-based network training and PIRL-based network training. © 2022 IEEE.

Energy-based Domain Adaption with Active Learning for Emerging Misinformation Detection

Classifying whether collected information related to emerging topics and domains is fake/incorrect is not an easy task because we do not have enough labeled data in the domains. Given labeled data from source domains (e.g., gossip and health) and limited labeled data from a newly emerging target domain (e.g., COVID-19 and Ukraine war), simply applying knowledge learned from source domains to the target domain may not work well because of different data distribution. To solve the problem, in this paper, we propose an energy-based domain adaptation with active learning for early misinformation detection. Given three real world news datasets, we evaluate our proposed model against two baselines in both domain adaptation and the whole pipeline. Our model outperforms the baselines, improving at least 5% in the domain adaptation task and 10% in the whole pipeline, showing effectiveness of our proposed approach. © 2022 IEEE.

ALPINE: Improving Remote Heart Rate Estimation using Contrastive Learning

Heart rate (HR) is a crucial physiological indicator of human health and can be used to detect cardiovascular disorders. The traditional HR estimation methods, such as electrocardiograms (ECG) and photoplethysmographs, require skin contact. Due to the increased risk of viral in- fection from skin contact, these approaches are avoided in the ongoing COVID-19 pandemic. Alternatively, one can use the non-contact HR estimation technique, remote photo- plethysmography (rPPG), wherein HR is estimated from the facial videos of a person. Unfortunately, the existing rPPG methods perform poorly in the presence of facial deformations. Recently, there has been a proliferation of deep learning networks for rPPG. However, these networks require large-scale labelled data for better generalization. To alleviate these shortcomings, we propose a method ALPINE, that is, A noveL rPPG technique for Improving the remote heart rate estimatioN using contrastive lEarning. ALPINE utilizes the contrastive learning framework during training to address the issue of limited labelled data and introduces diversity in the data samples for better network generalization. Additionally, we introduce a novel hybrid loss comprising contrastive loss, signal-to-noise ratio (SNR) loss and data fidelity loss. Our novel contrastive loss maximizes the similarity between the rPPG information from different facial regions, thereby minimizing the effect of local noise. The SNR loss improves the quality of temporal signals, and the data fidelity loss ensures that the correct rPPG signal is extracted. Our extensive experiments on publicly available datasets demonstrate that the proposed method, ALPINE outperforms the previous well-known rPPG methods. © 2023 IEEE.

Improving Semi-supervised Deep Learning by using Automatic Thresholding to Deal with Out of Distribution Data for COVID-19 Detection using Chest X-ray Images

Semi-supervised learning (SSL) leverages both labeled and unlabeled data for training models when the labeled data is limited and the unlabeled data is vast. Frequently, the unlabeled data is more widely available than the labeled data, hence this data is used to improve the level of generalization of a model when the labeled data is scarce. However, in real-world settings unlabeled data might depict a different distribution than the labeled dataset distribution. This is known as distribution mismatch. Such problem generally occurs when the source of unlabeled data is different from the labeled data. For instance, in the medical imaging domain, when training a COVID-19 detector using chest X-ray images, different unlabeled datasets sampled from different hospitals might be used. In this work, we propose an automatic thresholding method to filter out-of-distribution data in the unlabeled dataset. We use the Mahalanobis distance between the labeled and unlabeled datasets using the feature space built by a pre-trained Image-net Feature Extractor (FE) to score each unlabeled observation. We test two simple automatic thresholding methods in the context of training a COVID-19 detector using chest X-ray images. The tested methods provide an automatic manner to define what unlabeled data to preserve when training a semi-supervised deep learning architecture. © 2022 IEEE.

Fusing Deep Learning with Support Vector Machines to Detect COVID-19 in X-Ray Images

Deep neural networks are powerful learning machines that have laid foundations for most of the recent advancements in data analysis. Their most important advantage lies in learning how to extract the features from raw data, and these deep features are later classified with fully-connected layers. Although there exist more effective classifiers, including support vector machines, their high computational complexity is a serious obstacle in using them for classifying highly-dimensional and often huge datasets of deep features. We introduce a new framework which allows us to classify the deep features with evolutionarily-optimized support vector machines and we apply it to a real-life problem of detecting COVID-19 from X-ray images. We demonstrate that the proposed approach is highly effective and it outperforms well-established transfer learning strategies, thus improving the potential of existing pre-trained deep models. It can be particularly beneficial in cases when the amount and quality of labeled data is insufficient for performing full training of a network, but still too large for training a regular support vector machine. © 2022, The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd.

Classification of German Jungian Extraversion and Introversion Texts with Assessment of Changes during the COVID-19 Pandemic

The corona pandemic and countermeasures such as social distancing and lockdowns have confronted individuals with new challenges for their mental health and well-being. It can be assumed that the Jungian psychology types of extraverts and introverts react differently to these challenges. We propose a Bi-LSTM model with an attention mechanism for classifying introversion and extraversion from German tweets, which is trained on hand-labeled data created by 335 participants. With this work, we provide this novel dataset for free use and validation. The proposed model achieves solid performance with F1 = .72. Furthermore, we created a feature engineered logistic model tree (LMT) trained on hand-labeled tweets, to which the data is also made available with this work. With this second model, German tweets before and during the pandemic have been investigated. Extraverts display more positive emotions, whilst introverts show more insight and higher rates of anxiety. Even though such a model can not replace proper psychological diagnostics, it can help shed light on linguistic markers and to help understand introversion and extraversion better for a variety of applications and investigations. © European Language Resources Association (ELRA)

MDN: Meta-transfer Learning Method for Fake News Detection

The rapid development of social media has brought convenience to people’s lives, but at the same time, it has also led to the widespread and rapid dissemination of false information among the population, which has had a bad impact on society. Therefore, effective detection of fake news is of great significance. Traditional fake news detection methods require a large amount of labeled data for model training. For emerging events (such as COVID-19), it is often hard to collect high-quality labeled data required for training models in a short period of time. To solve the above problems, this paper proposes a fake news detection method MDN (Meta Detection Network) based on meta-transfer learning. This method can extract the text and image features of tweets to improve accuracy. On this basis, a meta-training method is proposed based on the model-agnostic meta-learning algorithm, so that the model can use the knowledge of different kinds of events, and can realize rapid detection on new events. Finally, it was trained on a multi-modal real data set. The experimental results show that the detection accuracy has reached 76.7%, the accuracy rate has reached 77.8%, and the recall rate has reached 85.3%, which is at a better level among the baseline methods. © 2022, Springer Nature Singapore Pte Ltd.

Self-supervised Learning for COVID-19 Detection from Chest X-ray Images

Most of existing computer vision applications rely on models trained on supervised corpora, this is contradictory to what the world is seeing with the explosion of massive sets of unlabeled data. In the field of medical imaging for example, creating labels is extremely time-consuming because professionals should spend countless hours looking at images to manually annotate, segment, etc. Recently, several works are looking for solutions to the challenge of learning effective visual representations with no human supervision. In this work, we investigate the potential of using a self-supervised learning as a pretraining phase in improving the classification of radiographic images when the amount of available annotated data is small. To do that, we propose to use a self-supervised framework by pretraining a deep encoder with contrastive learning on a chest X-ray dataset using no labels at all, and then fine-tuning it using only few labeled data samples. We experimentally demonstrate that an unsupervised pretraining on unlabeled data is able to learn useful representation from Chest X-ray images, and only few labeled data samples are sufficient to reach the same accuracy of a supervised model learnt on the whole annotated dataset. © 2022, Springer Nature Switzerland AG.

Measuring Shifts in Attitudes Towards COVID-19 Measures in Belgium

With the COVID-19 pandemic and subsequent measures in full swing, people voiced their opinions of these measures on social media. Although it remains an open problem to correctly interpret these voices and translate this to public policy, we work towards this by tracking support for corona-related measures in Belgium, a densely-populated trilingual country in Western Europe. To this end, we classify seven months' worth of Belgian COVID-related tweets using multilingual BERT and a manually labeled training set. The tweets are classified by which measure they refer to as well as by their stated opinion towards the curfew measure, for which we introduce a custom classification scheme (too strict, ok, too loose). Using this classification, we examine the change in topics discussed and views expressed over time and in reference to dates of related events such as the implementation of new measures or COVID-19 related announcements in the media. With these promising results, our contributions include (i) multiple multilingual BERT models trained on manually labeled data accompanied by (ii) historical analysis of the support for the curfew measure on Twitter and (iii) a thorough analysis of limitations and risks, together with best practices and a reference code book. © 2021 Kristen Scott, Pieter Delobelle, Bettina Berendt

Multilingual Sentiment Analysis on Short Text Document Using Semi-Supervised Machine Learning

Sentiment analysis is a task of identifying the sentiments in text which is often applied to analyzing text in social media, customer feedbacks, and product reviews. Various studies have explored how sentiment analysis can automatically done by using machine learning techniques. However, there has been few attempts in implementing sentiment analysis on multilingual text. Furthermore, most of the existing works uses labelled data to train and develop machine learning models for sentiment analysis. Using labelled data are often expensive and time consuming. In this study, a sentiment analysis model for multilingual text using semi-supervised machine learning was explored. The data used is composed of 50,788 tweets about COVID-19, these are cleaned by removing unnecessary characters, stop words, and emojis. After cleaning, the language of each tweet was identified, all tweets that are not written in Filipino or English were removed from the dataset. Afterwards, the tweets were all translated in English in preparation for the annotation phase. This study used an open-source tool, TextBlob, in annotating the tweets. TextBlob outputs the polarity of the text in vector representation. The TextBlob annotation were then validated by human experts through an inter-rater agreement. The level of agreement between the human annotations and TextBlob annotations have a substantial agreement with 0.78 Fleiss' Kappa value. Classifier models were developed using various machine learning algorithms. Based on the results of the experiment, SVC is the best performing model with count vectorizer as feature with an accuracy, precision, recall, and F1-score of 95%. For future work, fine tuning hyperparameters to optimize the models can be considered. © 2021 ACM.

Self-supervised Pretraining for Covid-19 and Other Pneumonia Detection from Chest X-ray Images

Artificial intelligence technology has made breakthroughs in computer vision and natural language processing in recent years. An important factor is that the technology analyzes tasks in a data-driven manner and automatically learns data representation from large representations of data sets for a special task. However, one of the challenges is the lacked enough labelled datasets for pneumonia detection from chest X-ray images, which usually has a small number of identically distributed labelled data for training and conflicts with data-driven deep learning. That also is the bottleneck of the development of medical imaging AI. To address this challenge, we propose a self-supervised pre-training method for Covid-19 and other pneumonia detection. The method includes pre-trained model training and transfer learning. The pre-trained model uses a self-supervised contrastive learning method to learn the general representations from source data with location-sensitive patches and multi-level features. Transfer learning includes three stages of training to specialize the representation from source data to target data. The experiments show that it has improved performance for Covid-19 detection and other pneumonia with few labelled data. © 2022, The Author(s), under exclusive license to Springer Nature Switzerland AG.