Personalized Machine Learning

Optimizing ML Performance and Design for Each Person.

There is a need, now more than ever, for personalized models and applications that can efficiently leverage big data and adapt to target individuals, especially in smart healthcare and medicine.

The main challenge in personalized modeling is how to effectively and intelligently extract the person information and embed it into machine learning algorithms to build personalized models that provide optimal performance (e.g., accuracy of mood prediction) for each individual in target (risk) group - thus, optimizing outcomes for target person rather than only the average group outcome.

To teach this new modeling paradigm, Dr Oggi Rudovic and Prof Roz Picard created a graduate course here at MIT Media Lab - MAS.S61: Personalized Machine Learning (PML).

For more details, see: courses.media.mit.edu/2017spring/mass61.