Machine Learning of Sleep and Wake Behaviors to Classify Self-Reported Evening Mood


The SNAPSHOT Study is a large-scale and long-term study that seeks to measure: Sleep, Networks, Affect, Performance, Stress, and Health using Objective Techniques. This study investigates: (1) how daily behaviors influence sleep, stress, mood, and other wellbeing-related factors (2) how accurately we can recognize/predict stress, mood and wellbeing (3) how interactions in a social network influence sleep behaviors. In this work we investigate the use of machine learning methods, using sleep and wake data, to predict mood. We seek to model behavioral patterns to predict these downturns in mood and begin to understand what will help build resilience to depression. Our results reveal that stress and happiness can be predicted most reliably from these signals, and that data collected while participants were asleep is particularly important to classifying happiness.

In Sleep
Natasha Jaques
Natasha Jaques

My research is focused on Social Reinforcement Learning–developing algorithms that use insights from social learning to improve AI agents’ learning, generalization, coordination, and human-AI interaction.