A5021812637- Machine Learning in Healthcare
- Artificial Intelligence in Healthcare and Education
- Healthcare cost, quality, practices
- Health Systems, Economic Evaluations, Quality of Life
- Topic Modeling
- Advanced Causal Inference Techniques
- Electronic Health Records Systems
- Contemporary Sociological Theory and Practice
- Artificial Intelligence in Healthcare
- Crime, Illicit Activities, and Governance
- Gothic Literature and Media Analysis
- Crime Patterns and Interventions
- Sepsis Diagnosis and Treatment
- Social and Cultural Dynamics
- Contemporary Literature and Criticism
- Ethics in Clinical Research
- Biomedical Text Mining and Ontologies
- Crime, Deviance, and Social Control
- Explainable Artificial Intelligence (XAI)
- Healthcare Operations and Scheduling Optimization
- Law in Society and Culture
- Insurance, Mortality, Demography, Risk Management
- Digital Games and Media
- Medical Coding and Health Information
- Domain Adaptation and Few-Shot Learning
Google (United States)
2022-2025
Boston College
1987-2025
Duke University
2024
DeepMind (United Kingdom)
2024
Stanford University
2018-2023
Stanford Medicine
2019-2022
Georgia Institute of Technology
2015-2018
Emory University
2015-2018
Abstract Large language models (LLMs) have demonstrated impressive capabilities, but the bar for clinical applications is high. Attempts to assess knowledge of typically rely on automated evaluations based limited benchmarks. Here, address these limitations, we present MultiMedQA, a benchmark combining six existing medical question answering datasets spanning professional medicine, research and consumer queries new dataset questions searched online, HealthSearchQA. We propose human...
In a provocative analysis written during the unfolding drama of 1992, Baudrillard draws on his concepts simulation and hyperreal to argue that Gulf War did not take place but was carefully scripted media event virtual war.Patton s introduction argues Baudrillard, more than any other critic War, correctly identified stakes involved in gestation New World Order.
Recent artificial intelligence (AI) systems have reached milestones in "grand challenges" ranging from Go to protein-folding. The capability retrieve medical knowledge, reason over it, and answer questions comparably physicians has long been viewed as one such grand challenge. Large language models (LLMs) catalyzed significant progress question answering; Med-PaLM was the first model exceed a "passing" score US Medical Licensing Examination (USMLE) style with of 67.2% on MedQA dataset....
Abstract Artificial intelligence as a medical device is increasingly being applied to healthcare for diagnosis, risk stratification and resource allocation. However, growing body of evidence has highlighted the algorithmic bias, which may perpetuate existing health inequity. This problem arises in part because systemic inequalities dataset curation, unequal opportunity participate research access. study aims explore standards, frameworks best practices ensuring adequate data diversity...
Large language models (LLMs) have shown promise in medical question answering, with Med-PaLM being the first to exceed a 'passing' score United States Medical Licensing Examination style questions. However, challenges remain long-form answering and handling real-world workflows. Here, we present 2, which bridges these gaps combination of base LLM improvements, domain fine-tuning new strategies for improving reasoning grounding through ensemble refinement chain retrieval. 2 scores up 86.5% on...
The SOD1 G93A mouse model of amyotrophic lateral sclerosis (ALS) is the most frequently used to examine ALS pathophysiology. There a lack homogeneity in usage mouse, including differences genetic background and gender, which could confound field's results.In an analysis 97 studies, we characterized progression for high transgene copy control on basis disease onset, overall lifespan, duration male female mice B6SJL C57BL/6J backgrounds quantified magnitudes between groups.Mean age at onset...
Temporal dataset shift associated with changes in healthcare over time is a barrier to deploying machine learning-based clinical decision support systems. Algorithms that learn robust models by estimating invariant properties across periods for domain generalization (DG) and unsupervised adaptation (UDA) might be suitable proactively mitigate shift. The objective was characterize the impact of temporal on prediction benchmark DG UDA algorithms improving model robustness. In this cohort...
Abstract Temporal distribution shift negatively impacts the performance of clinical prediction models over time. Pretraining foundation using self-supervised learning on electronic health records (EHR) may be effective in acquiring informative global patterns that can improve robustness task-specific models. The objective was to evaluate utility EHR improving in-distribution (ID) and out-of-distribution (OOD) Transformer- gated recurrent unit-based were pretrained up 1.8 M patients (382...
BackgroundArtificial intelligence (AI) has repeatedly been shown to encode historical inequities in healthcare. We aimed develop a framework quantitatively assess the performance equity of health AI technologies and illustrate its utility via case study.MethodsHere, we propose methodology whether prioritise for patient populations experiencing worse outcomes, that is complementary existing fairness metrics. developed Health Equity Assessment machine Learning (HEAL) designed four-step...
This symposium brings together the theory and practice of public sociology. The introduction sets out meanings sociology, emphasizing its plurality relation to multiple publics. From there, it frames sociology in policy, professional, critical sociologies. constellation division sociological labor varies over time between countries. We argue for a normative model antagonistic interdependence, which holds all four types equilibrium. core contains six autobiographical case studies from Boston...
Large language models (LLMs) have demonstrated impressive capabilities in natural understanding and generation, but the quality bar for medical clinical applications is high. Today, attempts to assess models' knowledge typically rely on automated evaluations limited benchmarks. There no standard evaluate model predictions reasoning across a breadth of tasks. To address this, we present MultiMedQA, benchmark combining six existing open question answering datasets spanning professional exams,...
Guidelines for the management of atherosclerotic cardiovascular disease (ASCVD) recommend use risk stratification models to identify patients most likely benefit from cholesterol-lowering and other therapies. These have differential performance across race gender groups with inconsistent behavior studies, potentially resulting in an inequitable distribution beneficial therapy. In this work, we leverage adversarial learning a large observational cohort extracted electronic health records...
The American College of Cardiology and the Heart Association guidelines on primary prevention atherosclerotic cardiovascular disease (ASCVD) recommend using 10-year ASCVD risk estimation models to initiate statin treatment. For guideline-concordant decision-making, estimates need be calibrated. However, existing are often miscalibrated for race, ethnicity sex based subgroups. This study evaluates two algorithmic fairness approaches adjust estimators (group recalibration equalised odds) their...
With growing application of machine learning (ML) technologies in healthcare, there have been calls for developing techniques to understand and mitigate biases these systems may exhibit.Fairness considerations the development ML-based solutions health particular implications Africa, which already faces inequitable power imbalances between Global North South.This paper seeks explore fairness global health, with Africa as a case study.We conduct scoping review propose axes disparities...
Introduction and Preliminaries Estimation Prediction Some Tests of Hypotheses Testing for Efficient Capital Markets