CyclOps: A Framework for Data Extraction, Model Evaluation and Drift Detection in Clinical Machine Learning

Explore the CyclOps framework for healthcare-oriented machine learning in this 30-minute conference talk from MLOps World: Machine Learning in Production. Learn about a unified approach to data extraction, model evaluation, and drift detection for clinical applications. Discover how CyclOps addresses challenges in healthcare ML, including standardizing data extraction from Electronic Health Records and building robust models that can handle dataset shifts. Gain insights into the framework's three main features: data querying and processing, baseline model training and evaluation, and dataset shift detection. Understand how CyclOps integrates with open-source components and provides Python APIs to empower interdisciplinary teams in preparing ML models for clinical deployment. Hear from speakers Amrit Krishnan and Vallijah Subasri as they discuss the framework's potential to improve decision-making in updating existing models within clinical settings.

CyclOps - A framework for Data Extraction, Model Evaluation and Drift Detection for Clinical Use