Lessons Learned from Online Chronic Disease Self-Management Program during COVID-19 Pandemic (preprint)

Background Chronic disease self-management education has been shown to be effective in improving self-efficacy and health outcomes. As a response to the COVID-19 pandemic, a fast and effective program adaptation and delivery was imperative for Yakima Valley Farm Workers Clinic from in-person to on-line to continue serving its patients and communities.Methods A retrospective self-administer questionnaire was developed for participants that attended the real time online program at high levels. Questionnaire was administered via SurveyMonkey. A total of 217 completed the questionnaire. Workshops were offered in English and Spanish to patients with any chronic condition(s). A1c data was captured using Epic.Results The peer-led and real time content-delivered online program demonstrated access capacity, preference for online workshops, and improved self-efficacy and outcomes among participants.Discussion Community health centers may benefit themselves and low-income communities by making real-time online content for chronic disease self-management programs available and deliverable to their patients.

Interpretable Reinforcement Learning via Neural Additive Models for Inventory Management (preprint)

The COVID-19 pandemic has highlighted the importance of supply chains and the role of digital management to react to dynamic changes in the environment. In this work, we focus on developing dynamic inventory ordering policies for a multi-echelon, i.e. multi-stage, supply chain. Traditional inventory optimization methods aim to determine a static reordering policy. Thus, these policies are not able to adjust to dynamic changes such as those observed during the COVID-19 crisis. On the other hand, conventional strategies offer the advantage of being interpretable, which is a crucial feature for supply chain managers in order to communicate decisions to their stakeholders. To address this limitation, we propose an interpretable reinforcement learning approach that aims to be as interpretable as the traditional static policies while being as flexible and environment-agnostic as other deep learning-based reinforcement learning solutions. We propose to use Neural Additive Models as an interpretable dynamic policy of a reinforcement learning agent, showing that this approach is competitive with a standard full connected policy. Finally, we use the interpretability property to gain insights into a complex ordering strategy for a simple, linear three-echelon inventory supply chain.

Machine Learning Model to Predict Vehicle Electrification Impacts on Urban Air Quality and Related Human Health Effects (preprint)

Air pollution is a prevailing environmental problem in cities worldwide. The future vehicle electrification (VE), which in Europe will be importantly fostered by the ban of thermal engines from 2035, is expected to have an important effect on urban air quality. Machine learning models represent an optimal tool for predicting changes in air pollutants concentrations in the context of future VE. For the city of Valencia (Spain), a XGBoost (eXtreme Gradient Boosting package) model was used in combination with SHAP (SHapley Additive exPlanations) analysis, both to investigate the importance of different factors explaining air pollution concentrations and predicting the effect of different levels of VE. The model was trained with 5 years of data including the COVID-19 lockdown period in 2020, in which mobility was strongly reduced resulting in unprecedent changes in air pollution concentrations. The interannual meteorological variability of 10 years was also considered in the analyses. For a 70% VE, the model predicted: 1) improvements in nitrogen dioxide pollution (-34% to -55% change in annual mean concentrations, for the different air quality stations), 2) a very limited effect on particulate matter concentrations (-1 to -4% change in annual means of PM2.5 and PM10), 3) heterogeneous responses in ground-level ozone concentrations (-3% to +12% change in the annual means of the daily maximum 8 h average concentrations). Even at a high VE increase of 70%, the 2021 World Health Organization Air Quality Guidelines will be exceeded for all pollutants in some stations. VE has a potentially important impact in terms of reducing NO2-associated premature mortality, but complementary strategies for reducing traffic and controlling all different air pollution sources should also be implemented to protect human health.

Journal Legal and Administrative Studies, Year XX, No. 2 (25) – 2021 (preprint)

In this issue of Journal of Legal and Administrative Studies (JLAS) are included scientific articles which debate problems from legal sciences field: OPEN STATEHOOD AND CONSTITUTIONAL CHANGE; ACTIVITIES OF PRIVATE MILITARY AND SECURITY COMPANIES IN SOME AFRICAN STATES; RATIO DECIDENDI END THE SCIENCE OF LAW; INITIAL TREATIES - PRIMARY SOURCES OF EUROPEAN UNION LAW; LIMITING SENIORS' RIGHT TO VOTE; THE CONCEPT OF CIVIL SERVICE INTEGRITY; THE TRANSPORT OF PERSONS UNDER NATIONAL LAW, IN THE CONTEXT OF THE CORONAVIRUS PANDEMIC etc.The Journal of Legal and Administrative Studies was founded in 2002 and it is dedicated to the academic teachers and researchers, lawyers, magistrates, Ph.D. students and post-doctoral researchers into legal and administrative sciences and their auxiliary sciences, from Romania and from abroad. Subjects for submission include the following main areas, but are not limited to them: Public Law, Private Law, Protection of human rights and protection against discrimination, European Union law, Forensics and Criminology, Legal Sociology, History of law, juridical philosophy.The journal promotes the original researches that contributes to the knowledge progress and are motivated by the necessity of studying the theory and practice in the mentioned areas. Also, the journal aims to create a forum for disciplinary and interdisciplinary debates and to become a standard in the national and international juridical and administrative research.

Predicting COVID-19 pandemic waves with biologically and behaviorally informed universal differential equations (preprint)

In the early stages of the COVID-19 pandemic, it became clear that pandemic waves and population responses were locked in a mutual feedback loop. The initial lull following strict interventions in the first wave often led to a second wave, as restrictions were relaxed. We test the ability of new hybrid machine learning techniques, namely universal differential equations (UDEs) with learning biases, to make predictions in a such a dynamic behavior-disease setting. We develop a UDE model for COVID-19 and test it both with and without learning biases describing simple assumptions about disease transmission and population response. Our results show that UDEs, particularly when supplied with learning biases, are capable of learning coupled behavior-disease dynamics and predicting second waves in a variety of populations. The model predicts a second wave of infections 55\% of the time across all populations, having been trained only on the first wave. The predicted second wave is larger than the first. Without learning biases, model predictions are hampered: the unbiased model predicts a second wave only 25\% of the time, typically smaller than the first. The biased model consistently predicts the expected increase in the transmission rate with rising mobility, whereas the unbiased model predicts a decrease in mobility as often as a continued increase. The biased model also achieves better accuracy on its training data thanks to fewer and less severely divergent trajectories. These results indicate that biologically informed machine learning can generate qualitatively correct mid to long-term predictions of COVID-19 pandemic waves.

Enhanced detection of the presence and severity of COVID-19 from CT scans using lung segmentation (preprint)

Improving automated analysis of medical imaging will provide clinicians more options in providing care for patients. The 2023 AI-enabled Medical Image Analysis Workshop and Covid-19 Diagnosis Competition (AI-MIA-COV19D) provides an opportunity to test and refine machine learning methods for detecting the presence and severity of COVID-19 in patients from CT scans. This paper presents version 2 of Cov3d, a deep learning model submitted in the 2022 competition. The model has been improved through a preprocessing step which segments the lungs in the CT scan and crops the input to this region. It results in a validation macro F1 score for predicting the presence of COVID-19 in the CT scans at 93.2% which is significantly above the baseline of 74\%. It gives a macro F1 score for predicting the severity of COVID-19 on the validation set for task 2 as 72.8% which is above the baseline of 38%.

Towards a Deep Learning Pain-Level Detection Deployment at UAE for Patient-Centric-Pain Management and Diagnosis Support: Framework and Performance Evaluation (preprint)

The outbreak of the COVID-19 pandemic revealed the criticality of timely intervention in a situation exacerbated by a shortage in medical staff and equipment. Pain-level screening is the initial step toward identifying the severity of patient conditions. Automatic recognition of state and feelings help in identifying patient symptoms to take immediate adequate action and providing a patient-centric medical plan tailored to a patient's state. In this paper, we propose a framework for pain-level detection for deployment in the United Arab Emirates and assess its performance using the most used approaches in the literature. Our results show that a deployment of a pain-level deep learning detection framework is promising in identifying the pain level accurately.

Prediction of SARS-CoV-2 transmission dynamics based on population-level cycle threshold values: A Machine learning and mechanistic modeling study (preprint)

Background: Polymerase chain reaction (PCR) cycle threshold (Ct) values can be used to estimate the viral burden of Severe Acute Respiratory Syndrome Coronavirus type 2 (SARS-CoV-2) and predict population-level epidemic trends. We investigated the use of machine learning (ML) and epidemic transmission modeling based on Ct value distribution for SARS-CoV-2 incidence prediction during an Omicron-predominant period. Methods: Using simulated data, we developed a ML model to predict the reproductive number based on Ct value distribution, and validated it on out-of-sample province-level data. We also developed an epidemiological model and fitted it to province-level data to accurately predict incidence. Results: Based on simulated data, the ML model predicted the reproductive number with highest performance on out-of-sample province-level data. The epidemiological model was validated on outbreak data, and fitted to province-level data, and accurately predicted incidence. Conclusions: These modeling approaches can complement traditional surveillance, especially when diagnostic testing practices change over time. The models can be tailored to different epidemiological settings and used in real time to guide public health interventions.

Optimal Machine Learning- and Deep Learning- driven algorithms for predicting the future value of investments: A systematic review and meta-analysis (preprint)

The COVID-19 pandemic and the increasing competitive landscape have led asset management companies to consider investing in applying Artificial Intelligence (AI)-driven technologies to minimise the risk and maximise the profitability of the investment funds they manage. Thus, a systematic review and a meta-analysis of the relevant literature were conducted to provide evidence-based informed recommendations on which AI-driven technologies could be leveraged for such purpose. Data on both Machine Learning (ML)- and Deep Learning (DL)-driven technologies applied to aid the management of investment funds in China and, specifically, in and around Shenzhen, were pooled from eleven eligible and recent studies (since 15 September 2017) and analysed accordingly. The key business-relevant and human-interpretable metrics representing their performance were identified in the root mean squared error (RMSE), in the same unit of currency of the investment funds, and the correlation strength between the predicted and actual values. One ML- and one DL-based algorithms were recommended to be used in the short and long terms respectively. In particular, the ML-based Gradient Boosting Decision Tree (GBDT) algorithm was found the most accurate in the relevant literature, e.g., 28.16% more accurate than the Support Vector Regressor (SVR), also having a highly competitive ability to capture trends in the actual values of investment funds (83.7% of correlation strength), whilst the Long-Short Term Memory (LSTM)-GBDT model was identified as the most accurate DL-based algorithm, 15.05% more accurate than the GBDT and with 13.2% higher ability to capture trends in the actual investment funds’ values than the GBDT.

In Search of Deep Learning Architectures for Load Forecasting: A Comparative Analysis and the Impact of the Covid-19 Pandemic on Model Performance (preprint)

In power grids, short-term load forecasting (STLF) is crucial as it contributes to the optimization of their reliability, emissions, and costs, while it enables the participation of energy companies in the energy market. STLF is a challenging task, due to the complex demand of active and reactive power from multiple types of electrical loads and their dependence on numerous exogenous variables. Amongst them, special circumstances, such as the COVID-19 pandemic, can often be the reason behind distribution shifts of load series. This work conducts a comparative study of Deep Learning (DL) architectures, namely Neural Basis Expansion Analysis Time Series Forecasting (N-BEATS), Long Short-Term Memory (LSTM), and Temporal Convolutional Networks (TCN), with respect to forecasting accuracy and training sustainability, meanwhile examining their out-of-distribution generalization capabilities during the COVID-19 pandemic era. A Pattern Sequence Forecasting (PSF) model is used as baseline. The case study focuses on day-ahead forecasts for the Portuguese national 15-minute resolution net load time series. The results can be leveraged by energy companies and network operators (i) to reinforce their forecasting toolkit with state-of-the-art DL models; (ii) to become aware of the serious consequences of crisis events on model performance; (iii) as a high-level model evaluation, deployment, and sustainability guide within a smart grid context.

Medical Face Masks and Emotion Recognition from the Body: Insights from a Deep Learning Perspective (preprint)

The COVID-19 pandemic has undoubtedly changed the standards and affected all aspects of our lives, especially social life. It has forced people to extensively wear medical face masks, in order to prevent transmission. This face occlusion can strongly irritate emotional reading from the face and urges us to incorporate the whole body for emotion recognition, as it needs to play a more major role, despite its complementary nature. In this paper, we want to conduct insightful studies about the effect of face occlusion on emotion recognition performance, and showcase the superiority of full body input over plain masked face. We utilize a deep learning model based on the Temporal Segment Network framework and aspire to fully overcome the consequences of the face mask. Although single RGB stream models can adapt and learn both facial and bodily features, this may lead to irrelevant information confusion. By processing those features separately and fusing their preliminary prediction scores with a late fusion scheme, we are more effectively taking advantage of both modalities. This architecture can also naturally support temporal modeling, by mingling information among neighboring segment frames. Experimental results suggest that spatial structure plays a more important role for an emotional expression, while temporal structure is complementary.

The Experiences of Online Students with Permanent Acquired Memory Related Issues: A qualitative study (preprint)

Higher education institutions within the United Kingdom have been making efforts to reduce the educational disparity that occurs between abled students and those with learning disabilities. Students with learning disabilities are more likely to abandon their studies than their able counterparts. However, given the unprecedented shift to online learning during the Covid − 19 pandemic, it is likely that those with learning disabilities have faced new challenges. Within psychology research into online learning, one population that has received little attention is students with permanent acquired memory-related issues. Therefore, this qualitative interpretative phenomenological analysis study explores the student experience of six mature university students who started online learning before Covid-19 to understand (i) what online learning means for students with permanent acquired memory related issues and (ii) what barriers and facilitators they encountered within their academic journey. Three themes were developed (i) Negotiating the challenges of online learning (ii) Online learning and the emotional experience, and (iii) Avoiding the gaps presented by online learning. The findings suggest that students’ online experience with permanent acquired memory loss was complex. Academic staff’s misunderstanding and the impact of Covid-19 combined to create several challenges, but support and individualised strategies aided course adherence.

Deep Learning and Medical Imaging for COVID-19 Diagnosis: A Comprehensive Survey (preprint)

COVID-19 (Coronavirus disease 2019) has been quickly spreading since its outbreak, impacting financial markets and healthcare systems globally. Countries all around the world have adopted a number of extraordinary steps to restrict the spreading virus, where early COVID-19 diagnosis is essential. Medical images such as X-ray images and Computed Tomography scans are becoming one of the main diagnostic tools to combat COVID-19 with the aid of deep learning-based systems. In this survey, we investigate the main contributions of deep learning applications using medical images in fighting against COVID-19 from the aspects of image classification, lesion localization, and severity quantification, and review different deep learning architectures and some image preprocessing techniques for achieving a preciser diagnosis. We also provide a summary of the X-ray and CT image datasets used in various studies for COVID-19 detection. The key difficulties and potential applications of deep learning in fighting against COVID-19 are finally discussed. This work summarizes the latest methods of deep learning using medical images to diagnose COVID-19, highlighting the challenges and inspiring more studies to keep utilizing the advantages of deep learning to combat COVID-19.

COVID-19 Infection Analysis Framework using Novel Boosted CNNs and Radiological Images (preprint)

COVID-19 is a new pathogen that first appeared in the human population at the end of 2019, and it can lead to novel variants of pneumonia after infection. COVID-19 is a rapidly spreading infectious disease that infects humans faster. Therefore, efficient diagnostic systems may accurately identify infected patients and thus help control their spread. In this regard, a new two-stage analysis framework is developed to analyze minute irregularities of COVID-19 infection. A novel detection Convolutional Neural Network (CNN), STM-BRNet, is developed that incorporates the Split-Transform-Merge (STM) block and channel boosting (CB) to identify COVID-19 infected CT slices in the first stage. Each STM block extracts boundary and region-smoothing-specific features for COVID-19 infection detection. Moreover, the various boosted channels are obtained by introducing the new CB and Transfer Learning (TL) concept in STM blocks to capture small illumination and texture variations of COVID-19-specific images. The COVID-19 CTs are provided with new SA-CB-BRSeg segmentation CNN for delineating infection in images in the second stage. SA-CB-BRSeg methodically utilized smoothening and heterogeneous operations in the encoder and decoder to capture simultaneously COVID-19 specific patterns that are region homogeneity, texture variation, and boundaries. Additionally, the new CB concept is introduced in the decoder of SA-CB-BRSeg by combining additional channels using TL to learn the low contrast region. The proposed STM-BRNet and SA-CB-BRSeg yield considerable achievement in accuracy: 98.01 %, Recall: 98.12%, F-score: 98.11%, and Dice Similarity: 96.396%, IOU: 98.845 % for the COVID-19 infectious region, respectively. The proposed two-stage framework significantly increased performance compared to single-phase and other reported systems and reduced the burden on the radiologists.

Discovering SARS-CoV-2 neoepitopes and the associated TCR-pMHC recognition mechanisms by combining single-cell sequencing, deep learning, and molecular dynamics simulation techniques (preprint)

The molecular mechanisms underlying the recognition of epitopes by T cell receptors (TCRs) are critical for activating T cell immune responses and rationally designing TCR-based therapeutics. Single-cell sequencing techniques vastly boost the accumulation of TCR sequences, while the limitation of available TCR-pMHC structures hampers further investigations. In this study, we proposed a comprehensive strategy that incorporates structural information and single-cell sequencing data to investigate the epitope-recognition mechanisms of TCRs. By antigen specificity clustering, we mapped the epitope sequences between epitope-known and epitope unknown TCRs from COVID-19 patients. One reported SARS-CoV-2 epitope, NQKLIANQF (S919-927), was identified for a TCR expressed by 614 T cells (TCR-614). Epitope screening also identified a potential cross-reactive epitope, KLKTLVATA (NSP31790-1798), for a TCR expressed by 204 T cells (TCR-204). According to the molecular dynamics (MD) simulations, we revealed the detailed epitope-recognition mechanisms for both TCRs. The structural motifs responsible for epitope recognition revealed by the MD simulations are consistent with the sequential features recognized by the sequence-based clustering method. This strategy will facilitate the discovery and optimization of TCR-based therapeutics. In addition, the comprehensive strategy can also promote the development of cancer vaccines in virtue of the ability to discover neoepitopes and epitope-recognition mechanisms.

Predicting Suspicious Money Laundering Transactions using Machine Learning Algorithms (preprint)

This study employs machine learning techniques to identify key drivers of suspicious activity reporting. The data for this study comes from all suspicious activities reported to the California government in 2018. In total, there were 45,000 records of data that represent various features. The paper uses linear regression along with Lasso, Ridge, and Elastic Net to perform feature regularization and address overfitting with the data. Other probabilistic and non-linear algorithms, namely, support vector machines, random forests, XGBoost, and CatBoost, were used to deal with the complexity of the data. The results from the  mean squared and root mean squared errors indicate that the ensemble tree-based algorithm performed better than the statistical and probabilistic models. The findings revealed that filings from regulators, the type of products, and customers' relationships with the institutions were the top contributors to SAR filings. Through the evaluation of a vast amount of data, this study provides valuable insights for identifying suspicious activities in financial transactions and has the potential to significantly improve suspicious transaction monitoring.

Development and validation of predictive models for mortality and length of hospitalisation in adults with community-onset bacteraemia under the COVID-19 pandemic: Application of early data dynamics (preprint)

Background Bacteraemia is associated with increased morbidity and mortality and contributes substantially to healthcare costs. The development of a scoring system to predict the short-term mortality and the length of hospitalisation (LOS) in patients with bacteraemia is essential to improve quality of care and reduce variance in hospital bed occupancy.Methods This multicentre study of patients hospitalised with community-onset bacteraemia retrospectively enrolled derivation and validation cohorts in the pre-COVID-19 and COVID-19 eras. This study developed five models to compare the performances of various scoring algorithms. Model I incorporated all variables available on day 0, Model II incorporated all variables available on day 3, and Models III, IV, and V incorporated the variables that changed from day 0 to day 3. This study adopted the statistical and machine learning (ML) methods to determine the crucial determinants of 30-day mortality and LOS in patients with community-onset bacteraemia, respectively.Results A total of 3,639 (81.4%) and 834 (18.6%) patients were included in the derivation and validation cohorts, respectively. Model IV best predicted 30-day mortality in both cohorts; it achieved the best performance (i.e., the largest area under the receiver operating characteristic [ROC] curve) according to the results of the logistic regression and most ML methods. The most frequently identified variables incorporated into Model IV were deteriorated consciousness from day 0 to day 3 and deteriorated respiration from day 0 to day 3. The generalised linear models and the majorities of ML methods also identified Model V as having the best performance (i.e., the lowest mean square error) in predicting LOS. The most frequently identified variables incorporated into Model V were deteriorated consciousness from day 0 to day 3, a body temperature ≤ 36.0°C or ≥ 39.0°C on day 3, and a diagnosis of complicated bacteraemia.Conclusions For hospitalised adults with community-onset bacteraemia, clinical variables that dynamically changed from day 0 to day 3 were crucial in predicting both the short-term mortality and their LOS.

A Benchmark Study by using various Machine Learning Models for Predicting Covid-19 trends (preprint)

Machine learning and deep learning play vital roles in predicting diseases in the medical field. Machine learning algorithms are widely classified as supervised, unsupervised, and reinforcement learning. This paper contains a detailed description of our experimental research work in that we used a supervised machine-learning algorithm to build our model for outbreaks of the novel Coronavirus that has spread over the whole world and caused many deaths, which is one of the most disastrous Pandemics in the history of the world. The people suffered physically and economically to survive in this lockdown. This work aims to understand better how machine learning, ensemble, and deep learning models work and are implemented in the real dataset. In our work, we are going to analyze the current trend or pattern of the coronavirus and then predict the further future of the covid-19 confirmed cases or new cases by training the past Covid-19 dataset by using the machine learning algorithm such as Linear Regression, Polynomial Regression, K-nearest neighbor, Decision Tree, Support Vector Machine and Random forest algorithm are used to train the model. The decision tree and the Random Forest algorithm perform better than SVR in this work. The performance of SVR and lasso regression are low in all prediction areas Because the SVR is challenging to separate the data using the hyperplane for this type of problem. So SVR mostly gives a lower performance in this problem. Ensemble (Voting, Bagging, and Stacking) and deep learning models(ANN) also predict well. After the prediction, we evaluated the model using MAE, MSE, RMSE, and MAPE. This work aims to find the trend/pattern of the covid-19.

Self-Supervised Curricular Deep Learning for Chest X-Ray Image Classification (preprint)

Deep learning technologies have already demonstrated a high potential to build diagnosis support systems from medical imaging data, such as Chest X-Ray images. However, the shortage of labeled data in the medical field represents one key obstacle to narrow down the performance gap with respect to applications in other image domains. In this work, we investigate the benefits of a curricular Self-Supervised Learning (SSL) pretraining scheme with respect to fully-supervised training regimes for pneumonia recognition on Chest X-Ray images of Covid-19 patients. We show that curricular SSL pretraining, which leverages unlabeled data, outperforms models trained from scratch, or pretrained on ImageNet, indicating the potential of performance gains by SSL pretraining on massive unlabeled datasets. Finally, we demonstrate that top-performing SSLpretrained models show a higher degree of attention in the lung regions, embodying models that may be more robust to possible external confounding factors in the training datasets, identified by previous works.

1D Convolutional neural networks and machine learning algorithms for spectral data classification with a case study for Covid-19 (preprint)

Machine and deep learning algorithms have increasingly been applied to solve problems in various areas of knowledge. Among these areas, Chemometrics has been benefited from the application of these algorithms in spectral data analysis. Commonly, algorithms such as Support Vector Machines and Partial Least Squares are applied to spectral datasets to perform classification and regression tasks. In this paper, we present a 1D convolutional neural networks (1D-CNN) to evaluate the effectiveness on spectral data obtained from spectroscopy. In most cases, the spectrum signals are noisy and present overlap among classes. Firstly, we perform extensive experiments including 1D-CNN compared to machine learning algorithms and standard algorithms used in Chemometrics on spectral data classification for the most known datasets available in the literature. Next, spectral samples of the SARS-COV2 virus, which causes the COVID-19, have recently been collected via spectroscopy was used as a case study. Experimental results indicate superior performance of 1D-CNN over machine learning algorithms and standard algorithms, obtaining an average accuracy of 96.5%, specificity of 98%, and sensitivity of 94%. The promissing obtained results indicate the feasibility to use 1D-CNN in automated systems to diagnose COVID-19 and other viral diseases in the future.