GTC 2020: AI Methods to Transfer Natural Language into Actionable Knowledge in Medicine: From Radiology, Pathology Reports to Social Media Posts
After clicking “Watch Now” you will be prompted to login or join.
Click “Watch Now” to login or join the NVIDIA Developer Program.
AI Methods to Transfer Natural Language into Actionable Knowledge in Medicine: From Radiology, Pathology Reports to Social Media Posts
Imon Banerjee, Emory University | Abeed Sarker, Emory University
Learn about the key types of use cases for AI-driven natural language processing (NLP) that will ultimately improve medical and public health practice by mining humongous amounts of free text from clinical notes and social media posts. First, we'll describe methods to leverage narrative reports associated with radiological scans to automatically generate labels for creating large annotated image datasets, and we'll highlight its application to different domains of radiology (CT, MR, US). Second, we'll discuss the challenge of clinical prediction and present innovative longitudinal XAI approaches (AI with explainability) to improve clinician acceptance. Finally, we'll discuss the emerging use of publicly available social media data in medicine and public health. While it is challenging to execute data mining from this resource, we'll present successful NLP pipelines that can convert this noisy language data into valuable and actionable knowledge.