Related references
Note: Only part of the references are listed.
Article
Biology
Mahesh Gour et al.
Summary: This study developed an uncertainty-aware neural network model, UA-ConvNet, for automated detection of COVID-19 disease from chest X-ray images with an estimation of uncertainty in model predictions. The model was evaluated on three different datasets and showed superiority in diagnosing COVID-19 cases from CXR images over existing methods.
COMPUTERS IN BIOLOGY AND MEDICINE
(2022)
Article
Multidisciplinary Sciences
Hamzeh Asgharnezhad et al.
Summary: This study focuses on uncertainty quantification of deep neural networks for COVID-19 detection in medical images. Through comprehensive experiments, it is shown that networks pertained on CXR images outperform networks pretrained on natural image datasets. Uncertainty quantification methods are capable of flagging risky predictions with high uncertainty estimates. Ensemble methods more reliably capture uncertainties. Previous studies mainly focused on generating single-valued predictions, while this paper introduces new performance metrics for the objective evaluation of uncertainty estimates.
SCIENTIFIC REPORTS
(2022)
Review
Statistics & Probability
Vincent Fortuin
Summary: The choice of prior is crucial for Bayesian deep learning models, and different models require different types of priors. Learners should pay more attention to prior specifications and gain inspiration from them.
INTERNATIONAL STATISTICAL REVIEW
(2022)
Article
Computer Science, Artificial Intelligence
Biraja Ghoshal et al.
Summary: Deep learning has made remarkable progress in medical image analysis, but current methods focus solely on accuracy of point predictions without considering the quality of outputs. This article proposes an uncertainty estimation framework, MC-DropWeights, to approximate Bayesian inference in DL by applying a Bernoulli distribution to model weights. By decomposing predictive probabilities into two main types of uncertainty and addressing mode collapse in variational inference, the MC-DropWeights method demonstrates improved estimation of uncertainty quality in image classification.
COMPUTATIONAL INTELLIGENCE
(2021)
Review
Statistics & Probability
Christopher Nemeth et al.
Summary: MCMC algorithms are considered the gold standard technique for Bayesian inference, but the computational cost can be prohibitive for large datasets, leading to the development of scalable Monte Carlo algorithms. One type of these algorithms is SGMCMC, which reduces per-iteration cost by utilizing data subsampling techniques.
JOURNAL OF THE AMERICAN STATISTICAL ASSOCIATION
(2021)
Review
Computer Science, Artificial Intelligence
Moloud Abdar et al.
Summary: Uncertainty quantification (UQ) methods are essential in reducing uncertainties in optimization and decision making processes. Bayesian approximation and ensemble learning techniques are widely used types of UQ methods.
INFORMATION FUSION
(2021)
Article
Computer Science, Artificial Intelligence
Gianluca Maguolo et al.
Summary: By manipulating X-Ray images, a more fair COVID-19 diagnostic testing protocol can be achieved. Neural networks may learn patterns in the dataset that are not relevant to COVID-19. Creating a fair testing protocol is a challenging task.
INFORMATION FUSION
(2021)
Review
Computer Science, Artificial Intelligence
Beatriz Garcia Santa Cruz et al.
Summary: The study systematically evaluated computer-aided diagnosis and stratification of COVID-19 based on chest X-ray, highlighting issues with bias assessment and quality control of datasets. Only a small number of datasets met criteria for proper risk bias assessment, with most datasets used in peer-reviewed papers having a high risk of bias.
MEDICAL IMAGE ANALYSIS
(2021)
Article
Computer Science, Artificial Intelligence
Ali Narin et al.
Summary: The 2019 novel coronavirus disease has rapidly spread worldwide, with nearly 101,917,147 cases reported, leading to a limited availability of COVID-19 test kits in hospitals. To address this, an automatic detection system using pre-trained convolutional neural network-based models was proposed for detecting coronavirus pneumonia-infected patients. Among the five models tested, ResNet50 demonstrated the highest accuracy in classifying COVID-19 patients from X-ray images.
PATTERN ANALYSIS AND APPLICATIONS
(2021)
Proceedings Paper
Computer Science, Interdisciplinary Applications
Biraja Ghoshal et al.
Summary: Reliable and cost-sensitive calibrated estimated uncertainty is crucial in many real-world applications where safety is critical and prediction problems are asymmetric. Bayesian decision theory provides a principled approach for optimal decision making under uncertainty.
2021 IEEE 34TH INTERNATIONAL SYMPOSIUM ON COMPUTER-BASED MEDICAL SYSTEMS (CBMS)
(2021)
Article
Computer Science, Information Systems
Matloob Khushi et al.
Summary: This study utilized 23 class imbalance methods and three classical classifiers to identify the best imbalance techniques for medical datasets. The results show that class imbalance learning can improve the classification ability of the model, with random forest performing the best in predictive ability among the over-sampling techniques used.
Article
Computer Science, Artificial Intelligence
Alex J. DeGrave et al.
Summary: Recent deep learning systems to detect COVID-19 from chest radiographs may rely on confounding factors rather than medical pathology, leading to accuracy issues when tested in new hospitals. The approach to obtain training data for these AI systems introduces a nearly ideal scenario for learning spurious shortcuts, raising concerns in medical-imaging AI. Evaluation of models on external data is insufficient to ensure reliance on medically relevant pathology, highlighting the importance of explainable AI for clinical deployment of machine-learning healthcare models.
NATURE MACHINE INTELLIGENCE
(2021)
Article
Computer Science, Information Systems
Hanan S. Alghamdi et al.
Summary: Chest X-ray imaging is a standard method for suspected COVID-19 cases, especially in resource-constrained areas. Artificial intelligence methods such as deep learning show promise for automatic diagnosis of COVID-19.
Article
Computer Science, Information Systems
Md. Milon Islam et al.
Summary: This paper provides an overview of systems developed for COVID-19 diagnosis using deep learning techniques, focusing on medical imaging modalities like CT and X-ray. It discusses well-known datasets used for training networks, data partitioning techniques, and performance measures. The paper concludes by addressing challenges and future trends in utilizing deep learning methods for COVID-19 detection. The aim is to aid experts and technicians in understanding and potentially further utilizing deep learning techniques in combating the pandemic.
Article
Multidisciplinary Sciences
Ana Rakita et al.
SCIENTIFIC REPORTS
(2020)
Article
Radiology, Nuclear Medicine & Medical Imaging
Keelin Murphy et al.
Article
Biology
Tanvir Mahmud et al.
COMPUTERS IN BIOLOGY AND MEDICINE
(2020)
Article
Computer Science, Information Systems
Muhammad E. H. Chowdhury et al.
Article
Computer Science, Information Systems
Julian D. Arias-Londono et al.
Review
Computer Science, Artificial Intelligence
Edmon Begoli et al.
NATURE MACHINE INTELLIGENCE
(2019)
Article
Mathematics, Interdisciplinary Applications
Yuling Yao et al.
Article
Computer Science, Theory & Methods
Aki Vehtari et al.
STATISTICS AND COMPUTING
(2017)
