A5030438721- COVID-19 diagnosis using AI
- Artificial Intelligence in Healthcare and Education
- Machine Learning in Healthcare
- Explainable Artificial Intelligence (XAI)
- Gestational Diabetes Research and Management
- Artificial Intelligence in Healthcare
- SARS-CoV-2 detection and testing
- Mobile Health and mHealth Applications
- Ethics and Social Impacts of AI
- Privacy-Preserving Technologies in Data
- Urinary Tract Infections Management
- Colorectal Cancer Screening and Detection
- Non-Invasive Vital Sign Monitoring
- Machine Learning and Data Classification
- Healthcare Operations and Scheduling Optimization
- Autopsy Techniques and Outcomes
- Bariatric Surgery and Outcomes
- Bacterial Identification and Susceptibility Testing
- Imbalanced Data Classification Techniques
- Heart Rate Variability and Autonomic Control
- Calcium signaling and nucleotide metabolism
- Ion channel regulation and function
- Viral Infections and Outbreaks Research
- Digital Imaging for Blood Diseases
- Vehicle Noise and Vibration Control
University of Oxford
2021-2024
Institute of Biomedical Science
2021-2024
Science Oxford
2023
Georgia State University
2010-2017
Kaiser Permanente San Jose Medical Center
2017
Abstract As patient health information is highly regulated due to privacy concerns, most machine learning (ML)-based healthcare studies are unable test on external cohorts, resulting in a gap between locally reported model performance and cross-site generalizability. Different approaches have been introduced for developing models across multiple clinical sites, however less attention has given adopting ready-made new settings. We introduce three methods do this—(1) applying “as-is” (2);...
Machine learning is becoming increasingly prominent in healthcare. Although its benefits are clear, growing attention being given to how these tools may exacerbate existing biases and disparities. In this study, we introduce an adversarial training framework that capable of mitigating have been acquired through data collection. We demonstrate proposed on the real-world task rapidly predicting COVID-19, focus site-specific (hospital) demographic (ethnicity) biases. Using statistical...
As models based on machine learning continue to be developed for healthcare applications, greater effort is needed ensure that these technologies do not reflect or exacerbate any unwanted discriminatory biases may present in the data. Here we introduce a reinforcement framework capable of mitigating have been acquired during data collection. In particular, evaluated our model task rapidly predicting COVID-19 patients presenting hospital emergency departments and aimed mitigate site...
Uncertainty in patients' COVID-19 status contributes to treatment delays, nosocomial transmission, and operational pressures hospitals. However, the typical turnaround time for laboratory PCR remains 12-24 h lateral flow devices (LFDs) have limited sensitivity. Previously, we shown that artificial intelligence-driven triage (CURIAL-1.0) can provide rapid screening using clinical data routinely available within 1 of arrival hospital. Here, aimed improve from emergency department availability...
Abstract With the rapid growth of memory and computing power, datasets are becoming increasingly complex imbalanced. This is especially severe in context clinical data, where there may be one rare event for many cases majority class. We introduce an imbalanced classification framework, based on reinforcement learning, training extremely data sets, extend it use multi-class settings. combine dueling double deep Q-learning architectures, formulate a custom reward function episode-training...
Multicentre training could reduce biases in medical artificial intelligence (AI); however, ethical, legal, and technical considerations can constrain the ability of hospitals to share data. Federated learning enables institutions participate algorithm development while retaining custody their data but uptake has been limited, possibly as deployment requires specialist software expertise at each site. We previously developed an intelligence-driven screening test for COVID-19 emergency...
Abstract Collaborative efforts in artificial intelligence (AI) are increasingly common between high-income countries (HICs) and low- to middle-income (LMICs). Given the resource limitations often encountered by LMICs, collaboration becomes crucial for pooling resources, expertise, knowledge. Despite apparent advantages, ensuring fairness equity of these collaborative models is essential, especially considering distinct differences LMIC HIC hospitals. In this study, we show that AI approaches...
Abstract Retinal fundus imaging is a powerful tool for disease screening and diagnosis in opthalmology. With the advent of machine learning artificial intelligence, particular modern computer vision classification algorithms, there broad scope technology to improve accuracy, increase accessibility reduce cost these processes. In this paper we present first deep model trained on Brazilian multi-label opthalmological datatset. We train classifier using over 16,000 clinically-labelled images....
Abstract Urinary tract infections are one of the most common bacterial worldwide; however, increasing antimicrobial resistance in pathogens is making it challenging for clinicians to correctly prescribe patients appropriate antibiotics. In this study, we present four interpretable machine learning-based decision support algorithms predicting resistance. Using electronic health record data from a large cohort diagnosed with potentially complicated UTIs, demonstrate high predictability...
Abstract Machine learning is becoming increasingly prominent in healthcare. Although its benefits are clear, growing attention being given to how machine may exacerbate existing biases and disparities. In this study, we introduce an adversarial training framework that capable of mitigating have been acquired through data collection or magnified during model development. For example, if one class over-presented errors/inconsistencies practice reflected the data, then a can be biased by these....
Abstract Background Tackling biases in medical artificial intelligence requires multi-centre collaboration, however, ethical, legal and entrustment considerations may restrict providers’ ability to participate. Federated learning (FL) eliminate the need for data sharing by allowing algorithm development across multiple hospitals without transfer. Previously, we have shown an AI-driven screening solution COVID-19 emergency departments using clinical routinely available within 1h of arrival...
Abstract The integration of artificial intelligence (AI) into healthcare systems within low-middle income countries (LMICs) has emerged as a central focus for various initiatives aiming to improve access and delivery quality. In contrast high-income (HICs), which often possess the resources infrastructure adopt innovative technologies, LMICs confront resource limitations such insufficient funding, outdated infrastructure, limited digital data, shortage technical expertise. Consequently, many...
Passive SR (sarcoplasmic reticulum) Ca2+ leak through the RyR (ryanodine receptor) plays a critical role in mechanisms that regulate [Ca2+]rest (intracellular resting myoplasmic free concentration) muscle. This process appears to be isoform-specific as expression of either RyR1 or RyR3 confers on myotubes different [Ca2+]rest. Using chimaeric RyR3-RyR1 receptors expressed dyspedic myotubes, we show isoform-dependent regulation is primarily defined by small region receptor encompassing amino...
Gestational diabetes mellitus (GDM) is often diagnosed during the last trimester of pregnancy, leaving only a short timeframe for intervention. However, appropriate assessment, management, and treatment have been shown to reduce complications GDM. This study introduces machine learning-based stratification system identifying patients at risk exhibiting high blood glucose levels, based on daily measurements electronic health record (EHR) data from GDM patients. We internally trained validated...
Abstract Collaborative efforts in artificial intelligence (AI) are increasingly common between high-income countries (HICs) and low-to middle-income (LMICs). Given the resource limitations often encountered by LMICs, collaboration becomes crucial for pooling resources, expertise, knowledge. Despite apparent advantages, ensuring fairness equity of these collaborative models is essential, especially considering distinct differences LMIC HIC hospitals. In this study, we show that AI approaches...
The analysis of extensive electronic health records (EHR) datasets often calls for automated solutions, with machine learning (ML) techniques, including deep (DL), taking a lead role. One common task involves categorizing EHR data into predefined groups. However, the vulnerability EHRs to noise and errors stemming from collection processes, as well potential human labeling errors, poses significant risk. This risk is particularly prominent during training DL models, where possibility...
In recent years, increasingly sophisticated computational and bioinformatics tools have evolved for the analyses of protein structure, function, ligand interactions, modeling energetics. This includes development algorithms to recursively evaluate side-chain rotamer permutations, identify regions in a 3D structure that meet some set search parameters, calculate minimize energy values, provide high-resolution visual theoretical modeling. Here we discuss interdependency between different areas...
Abstract As machine learning-based models continue to be developed for healthcare applications, greater effort is needed in ensuring that these technologies do not reflect or exacerbate any unwanted discriminatory biases may present the data. In this study, we introduce a reinforcement learning framework capable of mitigating have been acquired during data collection. particular, evaluated our model task rapidly predicting COVID-19 patients presenting hospital emergency departments, and...
Early detection of COVID-19 is an ongoing area research that can help with triage, monitoring and general health assessment potential patients may reduce operational strain on hospitals cope the coronavirus pandemic. Different machine learning techniques have been used in literature to detect cases using routine clinical data (blood tests, vital signs measurements). Data breaches information leakage when these models bring reputational damage cause legal issues for hospitals. In spite this,...
Abstract Healthcare data is highly sensitive and confidential, with strict regulations laws to protect patient privacy security. However, these impede the access of healthcare a wider AI research community. As result, often dominated by organisations larger datasets or limited silo-based development, where models are trained evaluated on population. Taking inspiration from non-sensitive nature summary statistics (mean, variance, etc.) data, this paper proposes geometrically-aggregated...
Abstract Gestational diabetes mellitus (GDM) is often diagnosed during the last trimester of pregnancy, leaving only a short timeframe for intervention. However, appropriate assessment, management, and treatment have been shown to reduce complications GDM. This study introduces machine learning-based stratification system identifying patients at risk exhibiting high blood glucose levels, based on daily measurements electronic health record (EHR) data from GDM patients. We internally trained...