Patient charts hold 80% of the information about a patient. Yet, this patient data is largely untapped today.
Understanding how to leverage this data leads to faster product innovation, improvements in patient care, and reduced cost.
In this session, Leo Janze, head of Data Engineering at Change Healthcare, will use real-life (de-identified) patient charts to share his experiences about:
– OCR, Optical Character Recognition
– Data labeling techniques using sparse data
– Semantics and sentence integrity
– Transformers, LSTM, and neural networks
– Data quality and bias detection
– Feature stores
After this session, you will have an appreciation for some common data challenges for AI in healthcare and get some ideas on how to address them with NLP.
Understanding how to leverage this data leads to faster product innovation, improvements in patient care, and reduced cost.