“Large language models allow the machine to essentially read through a large, large number of documents very quickly. So we're talking, thousands or millions of documents, in a very short amount of time, to find a needle in the haystack out of a large patient population.” Kevin Huang, Head of Data and Machine Learning at Notable, breaks down some of the most buzzworthy terms and complex concepts in the field of artificial intelligence and explains how each applies to healthcare.
In this episode, Kevin sits down for an in-depth conversation with Dr. Muthu Alagappan, chief medical officer at Notable. Among other things, the two discuss the field of artificial intelligence, how things like machine learning, deep learning, transformers, and large language models stand to impact the healthcare industry, and why now is the time for health systems to embrace technology solutions that are built around these latest technological advancements.
—-
Kevin Huang serves as Head of Data and Machine Learning for Notable, where he leads the development of the company’s intelligence platform, starting in the early days of natural language processing to work on incorporating large language models. Prior to Notable, Kevin was a data science technical lead at Change Healthcare. He holds a Ph.D. in electrical engineering from Stanford University, where he built computer simulations and new algorithms to study nanoscale LED devices.
—-
OUTLINE
Here are the timestamps for this episode.
(00:00) - Intro
(00:46) - What is artificial intelligence?
(03:00) - The differences between artificial intelligence and machine learning
(07:03) - Deep learning and natural language processing (NLP)
(11:53) - Limitations of NLP in healthcare
(14:49) - How do transformers connect NLP to large language models (LLMs)?
(18:43) - Defining large language models
(22:31) - Key AI terms summarized and defined
(23:51) - How do you create an LLM?
(25:44) - Limits to the power of LLMs
(29:55) - How to differentiate when everyone is using the same LLM
(33:33) - LLM use cases in healthcare
(39:18) - Safety, security, and LLMs
(44:18) - End
Relevant links