In this talk, we present our ongoing work utilizing more than 60 billion historical medical visits to create an automated layer for digital healthcare.
We will discuss the NLP challenges working with medical summaries in Hebrew.
We will present our Auto tagging ML model for automated entities extraction from medical summaries.
Our pipeline includes novelty deep models architectures built from scratch for sentence splitting, negation detection, entities relations and terms expansions.
We will share from our insights discovered from applying those systems in practice.