Inspiration Exchange: how can we make ML models as robust and useful as possible?

The van der Schaar Lab’s thirteenth Inspiration Exchange session took place virtually on December 2, 2021, and was attended by students and professionals from the AI and machine learning community.

This session raised a critically important question: how can we make ML models as robust and useful as possible?

Looking beyond predictions and analytics, there are a range of requirements and attributes that affect the degree of robustness and utility of a machine learning model—both in and beyond healthcare.

This session focused on three of these in particular: synthetic data, uncertainty estimation, and data imputation. Our lab members presented new and compelling work on each of these, while also conducting polls of the audience and asking attendees for their opinions and personal experiences.

Introduction – 0:00
Announcement regarding Ph.D. student recruitment – 3:32
Welcome message from Mihaela van der Schaar – 4:04
Presentation 1 by Boris van Breugel [synthetic data] – 5:38
Presentation 2 by Kamilė Stankevičiūtė [conformal time-series forecasting – uncertainty estimation] – 21:30
Presentation 3 by Trent Kyono [missing data imputation] – 37:45
Final discussion and wrap-up – 49:03
Intro to next sessions – 56:54

For an overview on synthetic data generation and assessment, click here. You can find more info about our work on uncertainty estimation here.

A full list of the lab’s publications can be found here.