A mask-guided attention deep learning model for COVID-19 diagnosis based on an integrated CT scan images database

The global extent of COVID-19 mutations and the consequent depletion of hospital resources highlighted the necessity of effective computer-assisted medical diagnosis. COVID-19 detection mediated by deep learning models can help diagnose this highly contagious disease and lower infectivity and mortality rates. Computed tomography (CT) is the preferred imaging modality for building automatic COVID-19 screening and diagnosis models. It is well-known that the training set size significantly impacts the performance and generalization of deep learning models. However, accessing a large dataset of CT scan images from an emerging disease like COVID-19 is challenging. Therefore, data efficiency becomes a significant factor in choosing a learning model. To this end, we present a multi-task learning approach, namely, a mask-guided attention (MGA) classifier, to improve the generalization and data efficiency of COVID-19 classification on lung CT scan images. The novelty of this method is compensating for the scarcity of data by employing more supervision with lesion masks, increasing the sensitivity of the model to COVID-19 manifestations, and helping both generalization and classification performance. Our proposed model achieves better overall performance than the single-task (without MGA module) baseline and state-of-the-art models, as measured by various popular metrics.

CoVEffect: interactive system for mining the effects of SARS-CoV-2 mutations and variants based on deep learning.

BACKGROUND: Literature about SARS-CoV-2 widely discusses the effects of variations that have spread in the past 3 years. Such information is dispersed in the texts of several research articles, hindering the possibility of practically integrating it with related datasets (e.g., millions of SARS-CoV-2 sequences available to the community). We aim to fill this gap, by mining literature abstracts to extract-for each variant/mutation-its related effects (in epidemiological, immunological, clinical, or viral kinetics terms) with labeled higher/lower levels in relation to the nonmutated virus. RESULTS: The proposed framework comprises (i) the provisioning of abstracts from a COVID-19-related big data corpus (CORD-19) and (ii) the identification of mutation/variant effects in abstracts using a GPT2-based prediction model. The above techniques enable the prediction of mutations/variants with their effects and levels in 2 distinct scenarios: (i) the batch annotation of the most relevant CORD-19 abstracts and (ii) the on-demand annotation of any user-selected CORD-19 abstract through the CoVEffect web application (http://gmql.eu/coveffect), which assists expert users with semiautomated data labeling. On the interface, users can inspect the predictions and correct them; user inputs can then extend the training dataset used by the prediction model. Our prototype model was trained through a carefully designed process, using a minimal and highly diversified pool of samples. CONCLUSIONS: The CoVEffect interface serves for the assisted annotation of abstracts, allowing the download of curated datasets for further use in data integration or analysis pipelines. The overall framework can be adapted to resolve similar unstructured-to-structured text translation tasks, which are typical of biomedical domains.

Interpretable clinical phenotypes among patients hospitalized with COVID-19 using cluster analysis.

Background: Multiple clinical phenotypes have been proposed for coronavirus disease (COVID-19), but few have used multimodal data. Using clinical and imaging data, we aimed to identify distinct clinical phenotypes in patients admitted with COVID-19 and to assess their clinical outcomes. Our secondary objective was to demonstrate the clinical applicability of this method by developing an interpretable model for phenotype assignment. Methods: We analyzed data from 547 patients hospitalized with COVID-19 at a Canadian academic hospital. We processed the data by applying a factor analysis of mixed data (FAMD) and compared four clustering algorithms: k-means, partitioning around medoids (PAM), and divisive and agglomerative hierarchical clustering. We used imaging data and 34 clinical variables collected within the first 24 h of admission to train our algorithm. We conducted a survival analysis to compare the clinical outcomes across phenotypes. With the data split into training and validation sets (75/25 ratio), we developed a decision-tree-based model to facilitate the interpretation and assignment of the observed phenotypes. Results: Agglomerative hierarchical clustering was the most robust algorithm. We identified three clinical phenotypes: 79 patients (14%) in Cluster 1, 275 patients (50%) in Cluster 2, and 203 (37%) in Cluster 3. Cluster 2 and Cluster 3 were both characterized by a low-risk respiratory and inflammatory profile but differed in terms of demographics. Compared with Cluster 3, Cluster 2 comprised older patients with more comorbidities. Cluster 1 represented the group with the most severe clinical presentation, as inferred by the highest rate of hypoxemia and the highest radiological burden. Intensive care unit (ICU) admission and mechanical ventilation risks were the highest in Cluster 1. Using only two to four decision rules, the classification and regression tree (CART) phenotype assignment model achieved an AUC of 84% (81.5-86.5%, 95 CI) on the validation set. Conclusions: We conducted a multidimensional phenotypic analysis of adult inpatients with COVID-19 and identified three distinct phenotypes associated with different clinical outcomes. We also demonstrated the clinical usability of this approach, as phenotypes can be accurately assigned using a simple decision tree. Further research is still needed to properly incorporate these phenotypes in the management of patients with COVID-19.

MepoGNN: Metapopulation Epidemic Forecasting with Graph Neural Networks

Epidemic prediction is a fundamental task for epidemic control and prevention. Many mechanistic models and deep learning models are built for this task. However, most mechanistic models have difficulty estimating the time/region-varying epidemiological parameters, while most deep learning models lack the guidance of epidemiological domain knowledge and interpretability of prediction results. In this study, we propose a novel hybrid model called MepoGNN for multi-step multi-region epidemic forecasting by incorporating Graph Neural Networks (GNNs) and graph learning mechanisms into Metapopulation SIR model. Our model can not only predict the number of confirmed cases but also explicitly learn the epidemiological parameters and the underlying epidemic propagation graph from heterogeneous data in an end-to-end manner. Experiment results demonstrate our model outperforms the existing mechanistic models and deep learning models by a large margin. Furthermore, the analysis on the learned parameters demonstrates the high reliability and interpretability of our model and helps better understanding of epidemic spread. Our model and data have already been public on GitHub https://github.com/deepkashiwa20/MepoGNN.git. © 2023, The Author(s), under exclusive license to Springer Nature Switzerland AG.

COVID-DenseNet: A Deep Learning Architecture to Detect COVID-19 from Chest Radiology Images

COVID-19 has a severe risk of spreading rapidly, the quick identification of which is essential. In this regard, chest radiology images have proven to be a practical screening approach for COVID-19 affected patients. This study proposes a deep learning-based approach using DenseNet-121 to detect COVID-19 patients effectively. We have trained and tested our model on the COVIDx dataset and performed both two-class and three-class classifications, achieving 96.49% and 93.71% accuracy, respectively. By successfully utilizing transfer learning, we achieve comparable performance to the state-of-the-art method while using 15 × fewer model parameters. Moreover, we performed an interpretability analysis using Grad-CAM to highlight the most significant image regions at test time. Finally, we developed a website that takes chest radiology images as input and detects the presence of COVID-19 or pneumonia and a heatmap highlighting the infected regions. Source code for reproducing results and model weights is available. © 2023, The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd.

Towards Providing Clinical Insights on Long Covid from Twitter Data

From the outset of the COVID-19 pandemic, social media has provided a platform for sharing and discussing experiences in real time. This rich source of information may also prove useful to researchers for uncovering evolving insights into post-acute sequelae of SARS-CoV-2 (PACS), commonly referred to as Long COVID. In order to leverage social media data, we propose using entity-extraction methods for providing clinical insights prior to defining subsequent downstream tasks. In this work, we address the gap between state-of-the-art entity recognition models and the extraction of clinically relevant entities which may be useful to provide explanations for gaining relevant insights from Twitter data. We then propose an approach to bridge the gap by utilizing existing configurable tools, and datasets to enhance the capabilities of these models. Code for this work is available at: https://github.com/VectorInstitute/ProjectLongCovid-NER. © 2023, The Author(s), under exclusive license to Springer Nature Switzerland AG.

Attention-based Multi-flow Network for COVID-19 Classification and Lesion Localization from Chest CT

COVID-19 has been rapidly spreading worldwide and infected more than 1 million people with over 690k deaths reported. It is urgent and crucial to identify COVID-19-infected patients by computed tomography (CT) accurately and rapidly. However, we found that two problems, weak supervision and lack of interpretability, hindered its development. To address these challenges, we propose an attention-based multi-flow network for COVID-19 classification and lesion localization from chest CT. In the proposed model, we built a Resnet-based multi-flow network to learn the local information and the longitudinal information from the full chest sequence slice. To assist doctors in decision-making, the attention mechanism integrated into the network, which can locate the key slices and key parts from a full chest CT sequence of patients. We have systematically evaluated our method on the CT images of 1031 cases, including 420 COVID-19 cases, 311CAP cases, and 300 non-pneumonia cases. Our method could obtain an average accuracy of 82.3%, with 85.7% sensitivity and 86.4 % specificity, which outperformed previous works. © 2022 IEEE.

SmartTensors: Unsupervised and Physics-informed Machine Learning Framework for the Geoscience Applications

SmartTensors (https://github.com/SmartTensors) is a novel framework for unsupervised and physics-informed machine learning for geoscience applications. The methods in SmartTensors AI platform are developed using advanced matrix/tensor factorization constrained by penalties enforcing robustness and interpretability (e.g., nonnegativity, sparsity, physics, and mathematical constraints;etc.). This framework has been applied to analyze diverse datasets related to a wide range of problems: from COVID-19 to wildfires and climate. Here, we will focus on the analysis of geothermal prospectivity of the Great Basin, U.S. The basin covers a vast area that is yet to be thoroughly explored to discover new geothermal resources. The available regional geochemical data are expected to provide critical information about the geothermal reservoir properties in the basin, including temperature, fluid/heat flow, boundary conditions, and spatial extent. The geochemical data may also include hidden (latent) information that is a proxy for geothermal prospectivity. We processed the sparse geochemical dataset of 18 geochemical attributes observed at 14,341 locations. The data are analyzed using our GeoThermalCloud toolbox for geothermal exploration (https://github.com/SmartTensors/GeoThermalCloud.jl) whichis also a part of the SmartTensors framework. An unsupervised machine learning using non-negative matrix factorization with customized k-means clustering (NMFk) as implemented in SmartTensors identified three hidden geothermal signatures representing low-, medium-, and high-temperature reservoirs, respectively (Fig). NMFk also evaluated the probability of occurrence of these types of resources through the studied region. NMFk also reconstructed attributes from sparse into continuous over the study domain. Future work will add in the ML analyses other regional- and site-scale datasets including geological, geophysical, and geothermal attributes. © 2022 Society of Exploration Geophysicists and the American Association of Petroleum Geologists.

Trustworthy Explanations for Knowledge Discovered from E-Health Records

In the current era of big data, very large amounts of data are generating at a rapid rate from a wide variety of rich data sources. Electronic health (e-health) records are examples of the big data. With the technological advancements, more healthcare practice has gradually been supported by electronic processes and communication. This enables health informatics, in which computer science meets the healthcare sector to address healthcare and medical problems. Embedded in the big data are valuable information and knowledge that can be discovered by data science, data mining and machine learning techniques. Many of these techniques apply "opaque box"approaches to make accurate predictions. However, these techniques may not be crystal clear to the users. As the users not necessarily be able to clearly view the entire knowledge discovery (e.g., prediction) process, they may not easily trust the discovered knowledge (e.g., predictions). Hence, in this paper, we present a system for providing trustworthy explanations for knowledge discovered from e-health records. Specifically, our system provides users with global explanations for the important features among the records. It also provides users with local explanations for a particular record. Evaluation results on real-life e-health records show the practicality of our system in providing trustworthy explanations to knowledge discovered (e.g., accurate predictions made). © 2022 IEEE.

Multi-objective optimization identifies a specific and interpretable COVID-19 host response signature.

The identification of a COVID-19 host response signature in blood can increase the understanding of SARS-CoV-2 pathogenesis and improve diagnostic tools. Applying a multi-objective optimization framework to both massive public and new multi-omics data, we identified a COVID-19 signature regulated at both transcriptional and epigenetic levels. We validated the signature's robustness in multiple independent COVID-19 cohorts. Using public data from 8,630 subjects and 53 conditions, we demonstrated no cross-reactivity with other viral and bacterial infections, COVID-19 comorbidities, or confounders. In contrast, previously reported COVID-19 signatures were associated with significant cross-reactivity. The signature's interpretation, based on cell-type deconvolution and single-cell data analysis, revealed prominent yet complementary roles for plasmablasts and memory T cells. Although the signal from plasmablasts mediated COVID-19 detection, the signal from memory T cells controlled against cross-reactivity with other viral infections. This framework identified a robust, interpretable COVID-19 signature and is broadly applicable in other disease contexts. A record of this paper's transparent peer review process is included in the supplemental information.

Towards Providing Clinical Insights on Long Covid from Twitter Data

From the outset of the COVID-19 pandemic, social media has provided a platform for sharing and discussing experiences in real time. This rich source of information may also prove useful to researchers for uncovering evolving insights into post-acute sequelae of SARS-CoV-2 (PACS), commonly referred to as Long COVID. In order to leverage social media data, we propose using entity-extraction methods for providing clinical insights prior to defining subsequent downstream tasks. In this work, we address the gap between state-of-the-art entity recognition models and the extraction of clinically relevant entities which may be useful to provide explanations for gaining relevant insights from Twitter data. We then propose an approach to bridge the gap by utilizing existing configurable tools, and datasets to enhance the capabilities of these models. Code for this work is available at: https://github.com/VectorInstitute/ProjectLongCovid-NER. © 2023, The Author(s), under exclusive license to Springer Nature Switzerland AG.

Towards Deep Interpretable Features

The problem of interpretability for binary image classification is considered through the lens of kernel two-sample tests and generative modelling. A feature extraction framework coined Deep Interpretable Features is developed, which is used in combination with IntroVAE, a generative model capable of high-resolution image synthesis. Experimental results on a variety of datasets, including COVID-19 chest x-rays demonstrate the benefits of combining deep generative models with the ideas from kernel-based hypothesis testing in moving towards more robust interpretable deep generative models.

Interpretable Vertebral Fracture Diagnosis

Do black-box neural network models learn clinically relevant features for fracture diagnosis? The answer not only establishes reliability, quenches scientific curiosity, but also leads to explainable and verbose findings that can assist the radiologists in the final and increase trust. This work identifies the concepts networks use for vertebral fracture diagnosis in CT images. This is achieved by associating concepts to neurons highly correlated with a specific diagnosis in the dataset. The concepts are either associated with neurons by radiologists pre-hoc or are visualized during a specific prediction and left for the user's interpretation. We evaluate which concepts lead to correct diagnosis and which concepts lead to false positives. The proposed frameworks and analysis pave the way for reliable and explainable vertebral fracture diagnosis. The code is publicly available (https://github.com/CAMP-eXplain-AI/Interpretable-Vertebral-Fracture-Diagnosis).

A multi-task FP-GNN framework enables accurate prediction of selective PARP inhibitors.

PARP (poly ADP-ribose polymerase) family is a crucial DNA repair enzyme that responds to DNA damage, regulates apoptosis, and maintains genome stability; therefore, PARP inhibitors represent a promising therapeutic strategy for the treatment of various human diseases including COVID-19. In this study, a multi-task FP-GNN (Fingerprint and Graph Neural Networks) deep learning framework was proposed to predict the inhibitory activity of molecules against four PARP isoforms (PARP-1, PARP-2, PARP-5A, and PARP-5B). Compared with baseline predictive models based on four conventional machine learning methods such as RF, SVM, XGBoost, and LR as well as six deep learning algorithms such as DNN, Attentive FP, MPNN, GAT, GCN, and D-MPNN, the evaluation results indicate that the multi-task FP-GNN method achieves the best performance with the highest average BA, F1, and AUC values of 0.753 ± 0.033, 0.910 ± 0.045, and 0.888 ± 0.016 for the test set. In addition, Y-scrambling testing successfully verified that the model was not results of chance correlation. More importantly, the interpretability of the multi-task FP-GNN model enabled the identification of key structural fragments associated with the inhibition of each PARP isoform. To facilitate the use of the multi-task FP-GNN model in the field, an online webserver called PARPi-Predict and its local version software were created to predict whether compounds bear potential inhibitory activity against PARPs, thereby contributing to design and discover better selective PARP inhibitors.

An interpretable forecasting framework for energy consumption and CO2 emissions

It is a well-established fact that energy consumption and production, as the primary sources of greenhouse gases, contribute to climate change and global warming issues. The analysis and estimation of the factors that contribute to these harmful gases will be of great assistance in the development of policies to reduce carbon dioxide emissions. In addition to identifying the factors related to energy consumption and CO2 emissions, forecasting the variable of interest as accurately as possible has a key role in increasing the efficiency of energy strategies to be implemented. Unlike studies in the literature, this study not only forecasts the future value of energy consumption and CO2 emissions but also determines the relationship between the predictions and the influential variables by revealing the contribution of each variable to the prediction. For this purpose, the study proposes an interpretable forecasting framework based on values of the Shapley additive explanation (SHAP) to provide a simpler explanation of machine learning (ML) models in forecasting energy consumption and CO2 emissions. The results obtained show that the total electricity generation from different energy sources is found to be the most important variable interacting positively with both energy consumption and CO2 emissions. Also, the influence of the predictors on projections made before and after COVID-19 has changed dramatically. The proposed method may assist policymakers in making future energy investments and establishing energy laws more accurately and efficiently as it explains the drivers of the forecasts.

A Systematic Approach for Contextualizing Focal Structure Analysis in Social Networks

Focal Structures are key sets of individuals who may be responsible for coordinating events, protests, or leading citizen engagement efforts on social media networks. Discovering focal structures that can promote online social campaigns is important but complex. Unlike influential individuals, focal structures can effect large-scale complex social processes. In our prior work, we applied a greedy algorithm and bi-level decomposition optimization solution to identify focal structures in social media networks. However, the outcomes lacked a contextual representation of the focal structures that affected interpretability. In this research, we present a novel Contextual Focal Structure Analysis (CFSA) model to enhance the discovery and the interpretability of the focal structures to provide the context in terms of the content shared by individuals in the focal structures through their communication network. The CFSA model utilizes multiplex networks, where the first layer is the users-users network based on mentions, replies, friends, and followers, and the second layer is the hashtag co-occurrence network. The two layers have interconnections based on the user hashtag relations. The model's performance was evaluated on real-world datasets from Twitter related to domestic extremist groups spreading information about COVID-19 and the Black Lives Matter (BLM) social movement during the 2020–2021 time. The model identified Contextual Focal Structure (CFS) sets revealing the context regarding individuals’ interests. We then evaluated the model's efficacy by measuring the influence of the CFS sets in the network using various network structural measures such as the modularity method, network stability, and average clustering coefficient values. The ranking Correlation Coefficient (RCC) was used to conduct a comparative evaluation with real-world scenarios. © 2022, The Author(s), under exclusive license to Springer Nature Switzerland AG.

Enhancing Online Epidemic Supervising System by Compartmental and GRU Fusion Model

The global pandemic, COVID-19, is an acute respiratory infectious disease caused by the 2019 novel coronavirus. Building the online epidemic supervising system to provide COVID-19 dynamic prediction and analysis has attracted the attention of the industry and applications community. In previous studies, the compartmental models and deep neural networks (DNNs) played important roles in predicting and analyzing the dynamics of the pandemic. Nevertheless, the compartmental model has limited ability to fit historical data and thus leads to unsatisfactory prediction accuracy due to the difficulty in parameter estimation. For DNNs, the lack of interpretability makes it difficult to explain the prediction results;thus, it cannot provide an in-depth understanding of the transmission mechanism of the pandemic. We propose a fusion model to leverage the merits of both models and resolve their shortcomings. The fusion model extracts epidemic-related knowledge from the state-of-the-art SEIDR compartmental model to guide the training of the GRU model, which can preserve the interpretability and achieve a good performance in predicting epidemic dynamics. This model can help to enhance the online epidemic supervising system by providing more accurate prediction results and deeper analysis. Our extensive experiments across multiple epidemic datasets from six European countries demonstrate that our model outperforms existing state-of-the-art baselines in predicting the active confirmed cases. More importantly, by analyzing the effective reproductive number, our method can reveal the risk of the second wave of the epidemic in Europe and justify the importance of social distancing to control the outbreak of the epidemic. © 2022 Junyi Ma et al.

Comparing Interpretable AI Approaches for the Clinical Environment: an Application to COVID-19

Machine Learning (ML) models play an important role in healthcare thanks to their remarkable performance in predicting complex phenomena. During the COVID-19 pandemic, different ML models were implemented to support decisions in the medical settings. However, clinical experts need to ensure that these models are valid, provide clinically useful information, and are implemented and used correctly. In this vein, they need to understand the logic behind the models to be able to trust them. Hence, developing transparent and interpretable models has increasing relevance. In this work, we applied four interpretable ML models including logistic regression, decision tree, pyFUME, and RIPPER to classify suspected COVID-19 patients based on clinical data collected from blood samples. After preprocessing the data set and training the models, we evaluate the models based on their predictive performance. Then, we illustrate that interpretability can be achieved in different ways. First, SHAP explanations are built from logistic regression and decision trees to obtain the features' importance. Then, the potential of pyFUME and RIPPER in providing inherent interpretability are reflected. Finally, potential ways to achieve trust in future studies are briefly discussed. © 2022 IEEE.

Using interpretable machine learning identify factors contributing to COVID-19 cases in the United States

COVID-19 has been declared as a “pandemic” by the World Health Organization (WHO) and has claimed more than a million lives and over 50 million confirmed cases worldwide as of 7th November 2020. This virus can be curbed in only two ways: vaccination and other by imposing non-pharmaceutical interventions (NPIs), which are behavioral changes to a person and community. Most of the nations worldwide have imposed NPIs in the form of social distancing and lockdowns, which have been effective in reducing the pace of the virus's spread, but continued implementation has deemed social and economic losses. Hence strategic implementation of NPIs in a burst of periods should be done based on educated decisions using data about population mobility trends to find hot zones that lead to a spike in cases. These decisions will positively impact the virus's spread with lower damage to social and economic aspects. © 2022 Elsevier Inc. All rights reserved.

A Neural Predictive Model of Negative Emotions for COVID-19

The long-lasting global pandemic of Coronavirus disease 2019 (COVID-19) has changed our daily life in many ways and put heavy burden on our mental health. Having a predictive model of negative emotions during COVID-19 is of great importance for identifying potential risky population. To establish a neural predictive model achieving both good interpretability and predictivity, we have utilized a large-scale (n =542) longitudinal dataset, alongside two independent samples for external validation. We built a predictive model based on psychologically meaningful resting state neural activities. The whole-brain resting-state neural activity and social-psychological profile of the subjects were obtained from Sept. to Dec. 2019 (Time 1). Their negative emotions were tracked and re-assessed twice, on Feb 22 (Time 2) and Apr 24 (Time 3), 2020, respectively. We first applied canonical correlation analysis on both the neural profiles and psychological profiles collected on Time 1, this step selects only the psychological meaningful neural patterns for later model construction. We then trained the neural predictive model using those identified features on data obtained on Time 2. It achieved a good prediction performance (r =0.44, p =8.13 ×10-27). The two most important neural predictors are associated with self-control and social interaction. This study established an effective neural prediction model of negative emotions, achieving good interpretability and predictivity. It will be useful for identifying potential risky population of emotional disorders related to COVID-19. IEEE