MIT data scientist and physician Leo Celi discusses the future of machine learning to make healthcare more equitable
By
Dominic Ali
MIT research scientist and staff physician with the Beth Israel Deaconess Medical Center.
On March 2, Dr. Leo Celi, a pioneer of artificial intelligence in medicine and a practicing intensive care unit physician, delivered a one-hour public lecture hosted by T-CAIREM as part of the Temerty Centre Speaker Series.
Celi discussed some of the current challenges of using machine learning in healthcare, and cautioned current and future researchers that data often carry implicit biases that affect patient outcomes.
"For those in this exciting field of medical artificial intelligence, we need to remind ourselves that the data we are using is collected in the process of care are heavily influenced by long-standing societal, cultural and institutional biases, as well as provider subjectivity in decision making," he told the assembled Zoom audience of 185.
"Before you even start exploring the data, and building your models, you have to look into how the data was put together," he explained.
Celi is best known for bringing together clinicians and data scientists to collect and analyze data routinely collected during the healthcare process. His team at MIT built and maintains the publicly available MIMIC database which provides healthcare data to thousands of users around the world.
To watch Dr. Celi’s lecture and discussion with T-CAIREM director Muhammad Mamdani, please visit T-CAIREM’s Past Events webpage.
