The role of artificial intelligence in pandemic responses: from epidemiological modeling to vaccine development

Molecular Biomedicine, Journal Year: 2025, Volume and Issue: 6(1)

Published: Jan. 3, 2025

Abstract Integrating Artificial Intelligence (AI) across numerous disciplines has transformed the worldwide landscape of pandemic response. This review investigates multidimensional role AI in pandemic, which arises as a global health crisis, and its preparedness responses, ranging from enhanced epidemiological modelling to acceleration vaccine development. The confluence technologies guided us new era data-driven decision-making, revolutionizing our ability anticipate, mitigate, treat infectious illnesses. begins by discussing impact on emerging countries worldwide, elaborating critical significance modelling, bringing enabling forecasting, mitigation response pandemic. In epidemiology, AI-driven models like SIR (Susceptible-Infectious-Recovered) SIS (Susceptible-Infectious-Susceptible) are applied predict spread disease, preventing outbreaks optimising distribution. also demonstrates how Machine Learning (ML) algorithms predictive analytics improve knowledge disease propagation patterns. collaborative aspect discovery clinical trials various vaccines is emphasised, focusing constructing AI-powered surveillance networks. Conclusively, presents comprehensive assessment impacts builds AI-enabled dynamic collaborating ML Deep (DL) techniques, develops implements trials. focuses screening, contact tracing monitoring virus-causing It advocates for sustained research, real-world implications, ethical application strategic integration strengthen collective face alleviate effects issues.

Language: Английский

---

MENet: A Mitscherlich function based ensemble of CNN models to classify lung cancer using CT scans

PLoS ONE, Journal Year: 2024, Volume and Issue: 19(3), P. e0298527 - e0298527

Published: March 11, 2024

Lung cancer is one of the leading causes cancer-related deaths worldwide. To reduce mortality rate, early detection and proper treatment should be ensured. Computer-aided diagnosis methods analyze different modalities medical images to increase diagnostic precision. In this paper, we propose an ensemble model, called Mitscherlich function-based Ensemble Network (MENet), which combines prediction probabilities obtained from three deep learning models, namely Xception, InceptionResNetV2, MobileNetV2, improve accuracy a lung model. The approach based on function, produces fuzzy rank combine outputs said base classifiers. proposed method trained tested two publicly available datasets, Iraq-Oncology Teaching Hospital/National Center for Cancer Diseases (IQ-OTH/NCCD) LIDC-IDRI, both these are computed tomography (CT) scan datasets. results in terms some standard metrics show that performs better than state-of-the-art methods. codes work at https://github.com/SuryaMajumder/MENet .

Language: Английский

---

Neural Networks for the Detection of COVID-19 and Other Diseases: Prospects and Challenges

Bioengineering, Journal Year: 2023, Volume and Issue: 10(7), P. 850 - 850

Published: July 18, 2023

Artificial neural networks (ANNs) ability to learn, correct errors, and transform a large amount of raw data into beneficial medical decisions for treatment care has increased in popularity enhanced patient safety quality care. Therefore, this paper reviews the critical role ANNs providing valuable insights patients’ healthcare efficient disease diagnosis. We study different types existing literature that advance ANNs’ adaptation complex applications. Specifically, we investigate advances predicting viral, cancer, skin, COVID-19 diseases. Furthermore, propose deep convolutional network (CNN) model called ConXNet, based on chest radiography images, improve detection accuracy disease. ConXNet is trained tested using image dataset obtained from Kaggle, achieving more than 97% 98% precision, which better other state-of-the-art models, such as DeTraC, U-Net, COVID MTNet, COVID-Net, having 93.1%, 94.10%, 84.76%, 90% 94%, 95%, 85%, 92% respectively. The results show performed significantly well relatively compared with aforementioned models. Moreover, reduces time complexity by dropout layers batch normalization techniques. Finally, highlight future research directions challenges, algorithms, insufficient available data, privacy security, integration biosensing ANNs. These require considerable attention improving scope diagnostic

Language: Английский

---

Physics-informed ensemble deep learning framework for improving state of charge estimation of lithium-ion batteries

Journal of Energy Storage, Journal Year: 2023, Volume and Issue: 73, P. 108915 - 108915

Published: Sept. 27, 2023

Language: Английский

---

SnapEnsemFS: a snapshot ensembling-based deep feature selection model for colorectal cancer histological analysis

Scientific Reports, Journal Year: 2023, Volume and Issue: 13(1)

Published: June 19, 2023

Abstract Colorectal cancer is the third most common type of diagnosed annually, and second leading cause death due to cancer. Early diagnosis this ailment vital for preventing tumours spread plan treatment possibly eradicate disease. However, population-wide screening stunted by requirement medical professionals analyse histological slides manually. Thus, an automated computer-aided detection (CAD) framework based on deep learning proposed in research that uses slide images predictions. Ensemble a popular strategy fusing salient properties several models make final such frameworks are computationally costly since it requires training multiple base learners. Instead, study, we adopt snapshot ensemble method, wherein, instead traditional method decision scores from snapshots Convolutional Neural Network (CNN) model, extract features penultimate layer CNN model. Since extracted same model but different environments, there may be redundancy feature set. To alleviate this, fed into Particle Swarm Optimization, meta-heuristic, dimensionality reduction space better classification. Upon evaluation publicly available colorectal histology dataset using five-fold cross-validation scheme, obtains highest accuracy 97.60% F1-Score 97.61%, outperforming existing state-of-the-art methods dataset. Further, qualitative investigation class activation maps provide visual explainability practitioners, as well justifies use CAD histology. Our source codes accessible at: https://github.com/soumitri2001/SnapEnsemFS .

Language: Английский

---

A Comparative Study of Machine Learning and Deep Learning Models for Automatic Parkinson’s Disease Detection from Electroencephalogram Signals

Diagnostics, Journal Year: 2025, Volume and Issue: 15(6), P. 773 - 773

Published: March 19, 2025

Background: Parkinson’s disease (PD) is one of the most prevalent, widespread, and intricate neurodegenerative disorders. According to experts, at least 1% people over age 60 are affected worldwide. In present time, early detection PD remains difficult due absence a clear consensus on its brain characterization. Therefore, there an urgent need for more reliable efficient technique PD. Using potential electroencephalogram (EEG) signals, this study introduces innovative method or classification patients through machine learning, as well accurate deep learning approach. Methods: We propose EEG-based approach by integrating advanced spectral feature engineering with models. (a) UC San Diego Resting State EEG dataset (b) IOWA dataset, we extract standardized from five key frequency bands—alpha, beta, theta, gamma, delta (α,β,θ,γ,δ) employ SVM (Support Vector Machine) classifier baseline, achieving notable accuracy. Furthermore, implement (CNN) complex multi-dimensional set combining power values all bands, which gives superior performance in distinguishing (both medication without states) healthy patients. Results: With five-fold cross-validation these two datasets, our approaches successfully achieve promising results subject dependent scenario. The achieves competitive accuracies 82% 94% (using gamma band) respectively non-PD subject. CNN classifier, model able capture major cross-frequency dependencies EEG; therefore, reach beyond 96% 99% those respectively. also perform experiments independent environment, where generates 68.09% Conclusions: Our findings, coupled extraction have provide non-invasive, efficient, diagnosing PD, further work aimed enhancing sets, inclusion large number subjects, improving generalizability across diverse environments.

Language: Английский

---

Pediatric chest X-ray diagnosis using neuromorphic models

Computers in Biology and Medicine, Journal Year: 2025, Volume and Issue: 192, P. 110173 - 110173

Published: April 23, 2025

This research presents an innovative neuromorphic method utilizing Spiking Neural Networks (SNNs) to analyze pediatric chest X-rays (PediCXR) identify prevalent thoracic illnesses. We incorporate spiking-based machine learning models such as Convolutional (SCNN), Residual (S-ResNet), and Hierarchical (HSNN), for radiographic analysis the publically available benchmark PediCXR dataset. These employ spatiotemporal feature extraction, residual connections, event-driven processing improve diagnostic precision. The HSNN model surpasses approaches from literature, with a classification accuracy of 96% across six illness categories, F1-score 0.95 specificity 1.0 in pneumonia detection. Our demonstrates that computing is feasible biologically inspired approach real-time medical imaging diagnostics, significantly improving performance.

Language: Английский

---

HyEpiSeiD: a hybrid convolutional neural network and gated recurrent unit model for epileptic seizure detection from electroencephalogram signals

Brain Informatics, Journal Year: 2024, Volume and Issue: 11(1)

Published: Aug. 21, 2024

Abstract Epileptic seizure (ES) detection is an active research area, that aims at patient-specific ES with high accuracy from electroencephalogram (EEG) signals. The early of crucial for timely medical intervention and prevention further injuries the patients. This work proposes a robust deep learning framework called HyEpiSeiD extracts self-trained features pre-processed EEG signals using hybrid combination convolutional neural network followed by two gated recurrent unit layers performs prediction based on those extracted features. proposed evaluated public datasets, UCI Epilepsy Mendeley datasets. model achieved 99.01% 97.50% classification accuracy, respectively, outperforming most state-of-the-art methods in epilepsy domain.

Language: Английский

---

Impact of image enhancement methods on lung disease diagnosis using x-ray images

International Journal of Information Technology, Journal Year: 2023, Volume and Issue: 15(7), P. 3521 - 3526

Published: Sept. 3, 2023

Language: Английский

---

Three-Stage Framework for Accurate Pediatric Chest X-ray Diagnosis Using Self-Supervision and Transfer Learning on Small Datasets

Diagnostics, Journal Year: 2024, Volume and Issue: 14(15), P. 1634 - 1634

Published: July 29, 2024

Pediatric respiratory disease diagnosis and subsequent treatment require accurate interpretable analysis. A chest X-ray is the most cost-effective rapid method for identifying monitoring various thoracic diseases in children. Recent developments self-supervised transfer learning have shown their potential medical imaging, including areas. In this article, we propose a three-stage framework with knowledge from adult X-rays to aid interpretation of pediatric thorax diseases. We conducted comprehensive experiments different pre-training fine-tuning strategies develop transformer or convolutional neural network models then evaluate them qualitatively quantitatively. The ViT-Base/16 model, fine-tuned CheXpert dataset, large emerged as effective, achieving mean AUC 0.761 (95% CI: 0.759–0.763) across six categories demonstrating high sensitivity (average 0.639) specificity 0.683), which are indicative its strong discriminative ability. baseline models, ViT-Small/16 ViT-Base/16, when directly trained on CXR only achieved scores 0.646 0.641–0.651) 0.654 0.648–0.660), respectively. Qualitatively, our model excels localizing diseased regions, outperforming pre-trained ImageNet other approaches, thus providing superior explanations. source code available online data can be obtained PhysioNet.

Language: Английский

---