Train, tune and distribute models with generative AI and machine learning capabilities
Summary
Those of us who had the good fortune to attend the Women in Data Science (WiDS) Conference at Stanford University on March 7, 2022 had a wonderful time. The conference was energizing and uplifting. You can find the full agenda here and listen to the recording for the whole day at the bottom of this blog entry. A recurring theme at the conference was data quality:
We were thrilled that Tanveer Syeda Mahmood, IBM Fellow, delivered a a technical vision talk (recording) and participated as a panelist for the data science in healthcare panel (recording) at the Conference. Among the stories that Tanveer shared was when she arrived in the US to take her PhD, she was told that she had arrived too late: AI was over: AI had all happened. Tanveer also told us that she was motivated to work healthcare when her father was misdiagnosed. You can read more about the impact of the misdiagnosis here The healthcare panel also focused on datasets and the need to work with experts and to understand the context and domain. A simple example was given where a family meeting at a hospital ICU was designated as a predictor of death, In general 80% of time is spent on understanding the data and 20% is sent on prediction. Issues concerning the need to mix labeled and unlabeled data, and to manage privacy and regulations in healthcare data.
LinkedIn Postings
Copy