Epitope-anchored contrastive transfer learning for paired CD8+ T cell receptor-antigen recognition (preprint)

Understanding the mechanisms of T-cell antigen recognition that underpin adaptive immune responses is critical for the development of vaccines, immunotherapies, and treatments against autoimmune diseases. Despite extensive research efforts, the accurate identification of T cell receptor (TCR)-antigen binding pairs remains a significant challenge due to the vast diversity and cross-reactivity of TCRs. Here, we propose a deep-learning framework termed Epitope-anchored Contrastive Transfer Learning (EPACT) tailored to paired human CD8+ TCRs from single-cell sequencing data. Harnessing the pre-trained representations and the contrastive co-embedding space, EPACT demonstrates state-of-the-art model generalizability in predicting TCR binding specificity for unseen epitopes and distinct TCR repertoires, offering potential values for practical outcomes in real-world scenarios. The contrastive learning paradigm achieves highly precise predictions for immunodominant epitopes and facilitates interpretable analysis of epitope-specific T cells. The TCR binding strength predicted by EPACT aligns well with the surge in spike-specific immune responses targeting SARS-CoV-2 epitopes after vaccination. We further fine-tune EPACT on TCR-epitope structural data to decipher the residue-level interactions involved in T-cell antigen recognition. EPACT not only exhibits superior capabilities in quantifying inter-chain distance matrices and identifying contact residue pairs but also corroborates the presence of molecular mimicry across multiple tumor-associated antigens. Together, EPACT can serve as a useful AI approach with significant potential in practical applications and contribute toward the development of TCR-based diagnostics and immunotherapies.

Mixing Individual and Collective Behaviours to Predict Out-of-Routine Mobility (preprint)

Predicting human displacements is crucial for addressing various societal challenges, including urban design, traffic congestion, epidemic management, and migration dynamics. While predictive models like deep learning and Markov models offer insights into individual mobility, they often struggle with out-of-routine behaviours. Our study introduces an approach that dynamically integrates individual and collective mobility behaviours, leveraging collective intelligence to enhance prediction accuracy. Evaluating the model on millions of privacy-preserving trajectories across three US cities, we demonstrate its superior performance in predicting out-of-routine mobility, surpassing even advanced deep learning methods. Spatial analysis highlights the model's effectiveness near urban areas with a high density of points of interest, where collective behaviours strongly influence mobility. During disruptive events like the COVID-19 pandemic, our model retains predictive capabilities, unlike individual-based models. By bridging the gap between individual and collective behaviours, our approach offers transparent and accurate predictions, crucial for addressing contemporary mobility challenges.

Discovery of RdRp Thumb-1 as a novel broad-spectrum antiviral family of targets and MDL-001 as a potent broad-spectrum inhibitor thereof - Part I: A Bioinformatics and Deep Learning Approach (preprint)

Positive sense, single-stranded RNA (+ssRNA) viruses consist of 12+ viral families that contain mild pathogens to pandemic-causing viruses like SARS-CoV-2, yet all share a vital and highly conserved RNA-dependent RNA polymerase (RdRp). While RdRp is the target of several viral inhibitors, the active site has several pitfalls when translating in vitro inhibitors to the clinic. The highly polar residues within the active site often necessitate the use of highly polar or charged compounds, especially when designing nucleoside analog inhibitors, posing significant challenges in optimizing drug-likeness and membrane permeability for clinical efficacy. Here, we investigated the broad-spectrum potential of the allosteric Thumb-1 cryptic site of the RdRp, which to date has only been adequately studied in Hepatitis C Virus (HCV). To explore this potential antiviral target, we used a suite of bioinformatics techniques, including homology modeling and multiple sequence alignments, to reveal the conserved landscape of the Thumb-1 site across +ssRNA viruses. We then used ChemPrint, our Mol-GDL (Molecular-Geometric Deep Learning) machine learning model to predict drug inhibition of the Thumb-1 site in RdRp across +ssRNA viruses. Here, we identify MDL-001 as a promising broad-spectrum antiviral candidate with favorable properties that enable oral and once-a-day dosing. We also show how the cryptic nature of the Thumb-1 site masks itself to conventional virtual screening techniques, like docking, where activity prediction is heavily based on solving or predicting an accurate structure of the open pocket. This study demonstrates the utility of this approach in drug discovery for broad-spectrum antivirals that target the Thumb-1 site.

A Closer Look at Spatial-Slice Features Learning for COVID-19 Detection (preprint)

Conventional Computed Tomography (CT) imaging recognition faces two significant challenges: (1) There is often considerable variability in the resolution and size of each CT scan, necessitating strict requirements for the input size and adaptability of models. (2) CT-scan contains large number of out-of-distribution (OOD) slices. The crucial features may only be present in specific spatial regions and slices of the entire CT scan. How can we effectively figure out where these are located? To deal with this, we introduce an enhanced Spatial-Slice Feature Learning (SSFL++) framework specifically designed for CT scan. It aim to filter out a OOD data within whole CT scan, enabling our to select crucial spatial-slice for analysis by reducing 70% redundancy totally. Meanwhile, we proposed Kernel-Density-based slice Sampling (KDS) method to improve the stability when training and inference stage, therefore speeding up the rate of convergence and boosting performance. As a result, the experiments demonstrate the promising performance of our model using a simple EfficientNet-2D (E2D) model, even with only 1% of the training data. The efficacy of our approach has been validated on the COVID-19-CT-DB datasets provided by the DEF-AI-MIA workshop, in conjunction with CVPR 2024. Our source code will be made available.

ViRNN: A Deep Learning Model for Viral Host Prediction (preprint)

Viral outbreaks are on the rise in the world, with the current outbreak of COVID-19 being among one of the worst thus far. Many of these outbreaks were the result of zoonotic transfer between species, and thus understanding and predicting the host of a virus is very important. With the rise of sequencing technologies it is becoming increasingly easy to sequence the full genomes of viruses, databases of publicly available viral genomes are widely available. We utilize a convolutional and recurrent neural network architecture (ViRNN) to predict the hosts for the Coronaviridae family (Coronaviruses) amongst the eleven most common hosts of this family. Our architecture performed with an overall accuracy of 90.55% on our test dataset, with a micro-average AUC-PR of 0.97. Performance was variable per host. ViRNN outperformed previously published methods like k-nearest neighbors and support vector machines, as well as previously published deep learning based methods. Saliency maps based on integrated gradients revealed a number of proteins in the viral genome that may be important interactions determining viral infection in hosts. Overall, this method provides an adaptable classifier capable of predicting host species from viral genomic sequence with high accuracy.

Artificial Intelligence (AI) Based Prediction of Mortality, for COVID-19 Patients (preprint)

For severely affected COVID-19 patients, it is crucial to identify high-risk patients and predict survival and need for intensive care (ICU). Most of the proposed models are not well reported making them less reproducible and prone to high risk of bias particularly in presence of imbalance data/class. In this study, the performances of nine machine and deep learning algorithms in combination with two widely used feature selection methods were investigated to predict last status representing mortality, ICU requirement, and ventilation days. Fivefold cross-validation was used for training and validation purposes. To minimize bias, the training and testing sets were split maintaining similar distributions. Only 10 out of 122 features were found to be useful in prediction modelling with Acute kidney injury during hospitalization feature being the most important one. The algorithms performances depend on feature numbers and data pre-processing techniques. LSTM performs the best in predicting last status and ICU requirement with 90%, 92%, 86% and 95% accuracy, sensitivity, specificity, and AUC respectively. DNN performs the best in predicting Ventilation days with 88% accuracy. Considering all the factors and limitations including absence of exact time point of clinical onset, LSTM with carefully selected features can accurately predict last status and ICU requirement. DNN performs the best in predicting Ventilation days. Appropriate machine learning algorithm with carefully selected features and balance data can accurately predict mortality, ICU requirement and ventilation support. Such model can be very useful in emergency and pandemic where prompt and precise

An early warning indicator trained on stochastic disease-spreading models with different noises (preprint)

The timely detection of disease outbreaks through reliable early warning signals (EWSs) is indispensable for effective public health mitigation strategies. Nevertheless, the intricate dynamics of real-world disease spread, often influenced by diverse sources of noise and limited data in the early stages of outbreaks, pose a significant challenge in developing reliable EWSs, as the performance of existing indicators varies with extrinsic and intrinsic noises. Here, we address the challenge of modeling disease when the measurements are corrupted by additive white noise, multiplicative environmental noise, and demographic noise into a standard epidemic mathematical model. To navigate the complexities introduced by these noise sources, we employ a deep learning algorithm that provides EWS in infectious disease outbreak by training on noise-induced disease-spreading models. The indicator's effectiveness is demonstrated through its application to real-world COVID-19 cases in Edmonton and simulated time series derived from diverse disease spread models affected by noise. Notably, the indicator captures an impending transition in a time series of disease outbreaks and outperforms existing indicators. This study contributes to advancing early warning capabilities by addressing the intricate dynamics inherent in real-world disease spread, presenting a promising avenue for enhancing public health preparedness and response efforts.

Exploring Machine Learning Strategies in COVID-19 Prognostic Modelling: A Systematic Analysis of Diagnosis, Classification and Outcome Prediction (preprint)

Background: The COVID-19 pandemic, which has impacted over 222 countries resulting in incalculable losses, has necessitated innovative solutions via machine learning (ML) to tackle the problem of overburdened healthcare systems. This study consolidates research employing ML models for COVID-19 prognosis, evaluates prevalent models and performance, and provides an overview of suitable models and features while offering recommendations for experimental protocols, reproducibility and integration of ML algorithms in clinical settings. Methods: We conducted a review following the PRISMA framework, examining ML utilisation for COVID-19 prediction. Five databases were searched for relevant studies up to 24 January 2023, resulting in 1,824 unique articles. Rigorous selection criteria led to 204 included studies. Top-performing features and models were extracted, with the area under the receiver operating characteristic curve (AUC) evaluation metric used for performance assessment. Results: This systematic review investigated 204 studies on ML models for COVID-19 prognosis across automated diagnosis (18.1%), severity classification (31.9%), and outcome prediction (50%). We identified thirty-four unique features in five categories and twenty-one distinct ML models in six categories. The most prevalent features were chest CT, chest radiographs, and advanced age, while the most frequently employed models were CNN, XGB, and RF. Top-performing models included neural networks (ANN, MLP, DNN), distance-based methods (kNN), ensemble methods (XGB), and regression models (PLS-DA), all exhibiting high AUC values. Conclusion: Machine learning models have shown considerable promise in improving COVID-19 diagnostic accuracy, risk stratification, and outcome prediction. Advancements in ML techniques and their integration with complementary technologies will be essential for expediting decision-making and informing clinical decisions, with long-lasting implications for healthcare systems globally.

Simple 2D Convolutional Neural Network-based Approach for COVID-19 Detection (preprint)

This study explores the use of deep learning techniques for analyzing lung Computed Tomography (CT) images. Classic deep learning approaches face challenges with varying slice counts and resolutions in CT images, a diversity arising from the utilization of assorted scanning equipment. Typically, predictions are made on single slices which are then combined for a comprehensive outcome. Yet, this method does not incorporate learning features specific to each slice, leading to a compromise in effectiveness. To address these challenges, we propose an advanced Spatial-Slice Feature Learning (SSFL++) framework specifically tailored for CT scans. It aims to filter out out-of-distribution (OOD) data within the entire CT scan, allowing us to select essential spatial-slice features for analysis by reducing data redundancy by 70\%. Additionally, we introduce a Kernel-Density-based slice Sampling (KDS) method to enhance stability during training and inference phases, thereby accelerating convergence and enhancing overall performance. Remarkably, our experiments reveal that our model achieves promising results with a simple EfficientNet-2D (E2D) model. The effectiveness of our approach is confirmed on the COVID-19-CT-DB datasets provided by the DEF-AI-MIA workshop.

Deep Learning-Based Approaches for Early Diagnosis and Tracking of COVID-19 (preprint)

The COVID-19 pandemic has prompted governments and specialists worldwide to collaborate in search of effective solutions and strategies for containment and eventual societal recovery. With advancements in equipment capabilities, remote communications, and on-board/distributed computing, information-based technologies are playing an increasingly critical role in identifying, detecting, and diagnosing potential COVID-19 cases. This study aims to explore factors for early diagnosis, tracking, and identification of COVID-19 spread, focusing on data collection and discussing opportunities for improvement. The study recognizes that deep learning models are well-suited for mitigating the impact of COVID-19, given the availability of a large volume of pandemic data through various technologies and collaborative efforts. While deep learning and big data approaches may not have been extensively implemented or clinically tested, they offer quick responses and valuable insights to medical staff and decision-makers. However, designing deep learning algorithms for COVID-19 presents numerous challenges. The quality and quantity of COVID-19 datasets need further improvement, demanding ongoing efforts from the research community to enhance data quality and reliability.

Pilot Randomised Controlled Trial - Nurture Early for Optimal Nutrition (NEON) Study: Community facilitator-led participatory learning and action (PLA) women's groups to improve infant feeding, care and dental hygiene practices in South Asian infants aged < 2 years in East London (preprint)

Background: The first 1000-days of life are a critical window and can result in adverse-health consequences due to inadequate nutrition. South-Asian (SA) communities face significant health-disparities, particularly in maternal and child-health. Community-based-interventions, often employing Participatory-Learning-and-Action (PLA) approaches, have effectively addressed health-inequalities in lower-income-nations. The aim of this study was to assess the feasibility of implementing a PLA-intervention to improve infant-feeding and care-practices in SA communities in London. Methods: Comprehensive-analyses were conducted to assess the feasibility/fidelity of this pilot-randomised-controlled-trial. Summary-statistics were computed to compare key-metrics (participant consent-rates, attendance, retention, intervention-support, perceived-effectiveness) against predefined-progression-rules guiding towards a definitive-trial. Secondary-outcomes were analysed, drawing insights from sources, such as The-Children's-Eating-Behaviour-Questionnaire (CEBQ), Parental-Feeding-Style-Questionnaires (PFSQ), 4-Day-Food-diary, and the Equality-Impact-Assessment (EIA) tool. Video-analysis of children's mealtime behaviour trends was conducted. Feedback-interviews were collected from participants. Results: Process-outcome measures met predefined-progression-rules for a definitive-trial which deemed the intervention as feasible. The secondary-outcomes analysis revealed no significant changes in children's BMI z-scores. This could be attributed to the abbreviated follow-up period of 6-months, reduced from 12-months, due to COVID-19-related delays. CEBQ analysis showed increased food-responsiveness, along with decreased emotional-over/undereating. A similar trend was observed in PFSQ. The EIA-tool found no potential discrimination areas, and video-analysis revealed a decrease in force-feeding-practices. Participant-feedbacks revealed improved awareness and knowledge-sharing. Conclusion: The study validates the feasibility of a community-oriented, co-adapted Participatory-Learning-and-Action approach for optimising infant-care among South-Asians in high-income countries. It underscores the potential of such interventions in promoting health-equity and improving health-outcomes. Further research is required to evaluate their wider impact.

Media Bias Matters: Understanding the Impact of Politically Biased News on Vaccine Attitudes in Social Media (preprint)

News media has been utilized as a political tool to stray from facts, presenting biased claims without evidence. Amid the COVID-19 pandemic, politically biased news (PBN) has significantly undermined public trust in vaccines, despite strong medical evidence supporting their efficacy. In this paper, we analyze: (i) how inherent vaccine stances subtly influence individuals' selection of news sources and participation in social media discussions; and (ii) the impact of exposure to PBN on users' attitudes toward vaccines. In doing so, we first curate a comprehensive dataset that connects PBN with related social media discourse. Utilizing advanced deep learning and causal inference techniques, we reveal distinct user behaviors between social media groups with various vaccine stances. Moreover, we observe that individuals with moderate stances, particularly the vaccine-hesitant majority, are more vulnerable to the influence of PBN compared to those with extreme views. Our findings provide critical insights to foster this line of research.

Scale up of COVID-19 testing services in NYC, 2020-2021: Lessons Learned to Maximize Reach, Equity and Timeliness (preprint)

During infectious disease epidemics, accurate diagnostic testing is key to rapidly identify and treat cases, and mitigate transmission.  When a novel pathogen is involved, building testing capacity and scaling testing services at the local level can present major challenges to healthcare systems, public health agencies and laboratories.  This mixed methods study examined lessons learned from the scale-up of SARS CoV-2 testing services in New York City (NYC), as a core part of NYC’s Test & Trace program. Using quantitative and geospatial analyses, the authors assessed program success at maximizing reach, equity and timeliness of SARS CoV-2 diagnostic testing services across NYC neighborhoods. Qualitative analysis of key informant interviews elucidated key decisions, facilitators and barriers involved in the scale-up of SARS-CoV-2 testing services. A major early facilitator was the ability to establish working relationships with private sector vendors and contractors to rapidly procure and manufacture necessary supplies locally.  NYC residents were, on average, less than 25 minutes away from free SARS CoV-2 diagnostic testing services by public transport, and services were successfully directed to most neighborhoods with highest transmission rates, with only one notable exception.   A key feature was to direct mobile testing vans and rapid antigen testing services to areas based on real-time neighborhood transmission data. Municipal leaders should prioritize fortifying supply chains, establish cross-sectoral partnerships to support and extend testing services, plan for continuous testing and validation of assays, ensure open communication feedback loops with CBO partners, and maintain infrastructure to support mobile services during infectious disease emergencies.

COVIDHealth: A Benchmark Twitter Dataset and Machine Learning based Web Application for Classifying COVID-19 Discussions (preprint)

The COVID-19 pandemic has had adverse effects on both physical and mental health. During this pandemic, numerous studies have focused on gaining insights into health-related perspectives from social media. In this study, our primary objective is to develop a machine learning-based web application for automatically classifying COVID-19-related discussions on social media. To achieve this, we label COVID-19-related Twitter data, provide benchmark classification results, and develop a web application. We collected data using the Twitter API and labeled a total of 6,667 tweets into five different classes: health risks, prevention, symptoms, transmission, and treatment. We extracted features using various feature extraction methods and applied them to seven different traditional machine learning algorithms, including Decision Tree, Random Forest, Stochastic Gradient Descent, Adaboost, K-Nearest Neighbour, Logistic Regression, and Linear SVC. Additionally, we used four deep learning algorithms: LSTM, CNN, RNN, and BERT, for classification. Overall, we achieved a maximum F1 score of 90.43% with the CNN algorithm in deep learning. The Linear SVC algorithm exhibited the highest F1 score at 86.13%, surpassing other traditional machine learning approaches. Our study not only contributes to the field of health-related data analysis but also provides a valuable resource in the form of a web-based tool for efficient data classification, which can aid in addressing public health challenges and increasing awareness during pandemics. We made the dataset and application publicly available, which can be downloaded from this link https://github.com/Bishal16/COVID19-Health-Related-Data-Classification-Website.

Developing a Multi-variate Prediction Model For COVID-19 From Crowd-sourced Respiratory Voice Data (preprint)

COVID-19 has affected more than 223 countries worldwide and in the Post-COVID Era, there is a pressing need for non-invasive, low-cost, and highly scalable solutions to detect COVID-19. We develop a deep learning model to identify COVID-19 from voice recording data. The novelty of this work is in the development of deep learning models for COVID-19 identification from only voice recordings. We use the Cambridge COVID-19 Sound database which contains 893 speech samples, crowd-sourced from 4352 participants via a COVID-19 Sounds app. Voice features including Mel-spectrograms and Mel-frequency cepstral coefficients (MFCC) and CNN Encoder features are extracted. Based on the voice data, we develop deep learning classification models to detect COVID-19 cases. These models include Long Short-Term Memory (LSTM) and Convolutional Neural Network (CNN) and Hidden-Unit BERT (HuBERT). We compare their predictive power to baseline machine learning models. HuBERT achieves the highest accuracy of 86\% and the highest AUC of 0.93. The results achieved with the proposed models suggest promising results in COVID-19 diagnosis from voice recordings when compared to the results obtained from the state-of-the-art.

Reconstructing Masked Faces using Variational Quantized Variational Auto Encoders and Recognition using DCNN-ELM Hybrid Framework (preprint)

The reconstruction of the face has historically been a significant issue in medical and forensic science. The presence of COVID-19 has added a significant new dimension. To model a new face, plastic surgery and informatics are employed, representing cyber forensics with challenges. The classic facial recognition techniques suffer from major drawbacks when face masks are widely used. As a result, new techniques are now being tried and tested to reconstruct a face from a collection of masked facial images. To determine the identification accuracy and other parameters/metrics, these faces are compared to real-world images of the same subject. Our research focuses on the task of post-mask face reconstruction, addressing the pressing need for precise and reliable techniques. We evaluate the effectiveness of three key algorithms: Edge Connect, Gated Convolution, and Hierarchical Variational Vector Quantized Autoencoders (HVQVAE). We use two synthetic datasets, MaskedFace-CelebA and MaskedFace-CelebAHQ, to rigorously assess the quality of reconstructed faces using metrics such as PSNR, SSIM, UIQI, and NCORR. Gated Convolution (GC) emerges as the superior choice in terms of image quality. To validate our findings, we employ five classifiers (Vgg16, Vgg19, ResNet50, ResNet101, ResNET152) and explore Extreme Learning Machine (ELM) and Support Vector Machine (SVM) as novel approaches for face recognition. A comprehensive ablation study reinforces our conclusion that Generative Convolution (GC) excels among the three models. Our research offers valuable insights into face reconstruction amid widespread mask usage, emphasizing innovative methodologies to address contemporary challenges in the field.

Multiple Instance Learning for Cheating Detection and Localization in Online Examinations (preprint)

The spread of the Coronavirus disease-2019 epidemic has caused many courses and exams to be conducted online. The cheating behavior detection model in examination invigilation systems plays a pivotal role in guaranteeing the equality of long-distance examinations. However, cheating behavior is rare, and most researchers do not comprehensively take into account features such as head posture, gaze angle, body posture, and background information in the task of cheating behavior detection. In this paper, we develop and present CHEESE, a CHEating detection framework via multiplE inStancE learning. The framework consists of a label generator that implements weak supervision and a feature encoder to learn discriminative features. In addition, the framework combines body posture and background features extracted by 3D convolution with eye gaze, head posture and facial features captured by OpenFace 2.0. These features are fed into the spatio-temporal graph module by stitching to analyze the spatio-temporal changes in video clips to detect the cheating behaviors. Our experiments on three datasets, UCF-Crime, ShanghaiTech and Online Exam Proctoring (OEP), prove the effectiveness of our method as compared to the state-of-the-art approaches, and obtain the frame-level AUC score of 87.58% on the OEP dataset.

Online early supportive psychological intervention for bereaved families during COVID-19: Lessons Learned and Implementation Challenges (preprint)

As the COVID-19 pandemic has created complex conditions and the horrific loss of numerous lives, grieving the loss of loved ones in close families can be extremely difficult. To reduce the suffering of the loss and prevent the development of complicated grief, it is necessary to provide bereavement care as soon as possible. Therefore, we quickly developed a complete online program that included supportive psycho-therapeutic interventions and psychiatric counseling. The structure of all services is the main emphasis of the study, which also emphasizes the quantitative components and the unique characteristics of the interventions. Based on the lesson learned, we discussed the difficulties experienced in putting into practice an internet-based preventive service.

Mitigating Machine Learning Bias Between High Income and Low-Middle Income Countries for Enhanced Model Fairness and Generalizability (preprint)

Collaborative efforts in artificial intelligence (AI) are increasingly common between high-income countries (HICs) and low- to middle-income countries (LMICs). Given the resource limitations often encountered by LMICs, collaboration becomes crucial for pooling resources, expertise, and knowledge. Despite the apparent advantages, ensuring the fairness and equity of these collaborative models is essential, especially considering the distinct differences between LMIC and HIC hospitals. In this study, we show that collaborative AI approaches can lead to divergent performance outcomes across HIC and LMIC settings, particularly in the presence of data imbalances. Through a real-world COVID-19 screening case study, we demonstrate that implementing algorithmic-level bias mitigation methods significantly improves outcome fairness between HIC and LMIC sites while maintaining high diagnostic sensitivity. We compare our results against previous benchmarks, utilizing datasets from four independent United Kingdom Hospitals and one Vietnamese hospital, representing HIC and LMIC settings, respectively.

Epidemic Modeling using Hybrid of Time-varying SIRD, Particle Swarm Optimization, and Deep Learning (preprint)

Epidemiological models are best suitable to model an epidemic if the spread pattern is stationary. To deal with non-stationary patterns and multiple waves of an epidemic, we develop a hybrid model encompassing epidemic modeling, particle swarm optimization, and deep learning. The model mainly caters to three objectives for better prediction: 1. Periodic estimation of the model parameters. 2. Incorporating impact of all the aspects using data fitting and parameter optimization 3. Deep learning based prediction of the model parameters. In our model, we use a system of ordinary differential equations (ODEs) for Susceptible-Infected-Recovered-Dead (SIRD) epidemic modeling, Particle Swarm Optimization (PSO) for model parameter optimization, and stacked-LSTM for forecasting the model parameters. Initial or one time estimation of model parameters is not able to model multiple waves of an epidemic. So, we estimate the model parameters periodically (weekly). We use PSO to identify the optimum values of the model parameters. We next train the stacked-LSTM on the optimized parameters, and perform forecasting of the model parameters for upcoming four weeks. Further, we fed the LSTM forecasted parameters into the SIRD model to forecast the number of COVID-19 cases. We evaluate the model for highly affected three countries namely; the USA, India, and the UK. The proposed hybrid model is able to deal with multiple waves, and has outperformed existing methods on all the three datasets.