Get our free extension to see links to code for papers anywhere online!

Chrome logo Add to Chrome

Firefox logo Add to Firefox


Multimodal Privacy-preserving Mood Prediction from Mobile Data: A Preliminary Study

Dec 04, 2020
Terrance Liu, Paul Pu Liang, Michal Muszynski, Ryo Ishii, David Brent, Randy Auerbach, Nicholas Allen, Louis-Philippe Morency



Mental health conditions remain under-diagnosed even in countries with common access to advanced medical care. The ability to accurately and efficiently predict mood from easily collectible data has several important implications towards the early detection and intervention of mental health disorders. One promising data source to help monitor human behavior is from daily smartphone usage. However, care must be taken to summarize behaviors without identifying the user through personal (e.g., personally identifiable information) or protected attributes (e.g., race, gender). In this paper, we study behavioral markers or daily mood using a recent dataset of mobile behaviors from high-risk adolescent populations. Using computational models, we find that multimodal modeling of both text and app usage features is highly predictive of daily mood over each modality alone. Furthermore, we evaluate approaches that reliably obfuscate user identity while remaining predictive of daily mood. By combining multimodal representations with privacy-preserving learning, we are able to push forward the performance-privacy frontier as compared to unimodal approaches.



Share this with someone who'll enjoy it:

   Access Paper Source



Share this with someone who'll enjoy it: