Using the power of Azure and AI to help fight the COVID pandemic
summary
A reproducible pipeline processes the CORD‑19 corpus (~800,000 papers, ~30,000/month), extracts medical entities mapped to UMLS, and stores JSON records in Cosmos DB for querying and visualization in Power BI. Models use PubMed BERT or Microsoft’s Text Analytics for Health; training on small labeled sets (e.g. BC5‑CDR, ~300 entities) is feasible in hours on one GPU. Orchestration and parallelization run on Azure Machine Learning. Outcome: searchable dosages, co‑occurrence graphs, and temporal spikes (hydroxychloroquine rise then fall).
AI-written summary · about
chapters (9)
- 0:00Start
- 2:49What Nlp Tasks Can Do
- 6:00Extract Entities and Relations
- 11:01Jupiter Notebook
- 14:31Schedule the Experiment
- 15:13Command Line Approach
- 20:04Install the Azure Text Analytics Sdk
- 22:09Cosmos Db
- 29:30Power Bi
related talks
57mHow Machine Learning and AI Can Support the Fight Against COVID-19
55mTake Control of the Data of You
01h 00mTake Control of the Data of You
12mEmpowering digital transformation with .NET, ML and Azure
13mBuilding High-Performance AI Literature Processing
33mBuild your own advanced AI solutions on Azure Databricks
37mA novel way to manage and represent unstructured clinical data
46mFrom Data to Insights: Building AI-Powered Data APIs