Psychological processes move fast. One study by Stanford University found people switch tasks on digital devices every 19 seconds. The Screenomics is a new research paradigm that aims to collect screenshots of smartphones to study psychological processes comprehensively. 


As a member of the Screenomics team, I study how to use artificial intelligence to identify key predictors for the behavior of interest. My work found the temporal features of visual stimulation is the most important predictor of task switching. This finding indicates the sequence of features on screenshots (color, text, sentiment), not a single screenshot alone, can explain the task-switching behavior to maintain engagement or switch to other content. 

For the next step, I am interested in exploring the possibilities of using inverse reinforcement learning to quantify the reward structure of features on the screen. This is an exciting area of method development in behavior modeling. 

Related Work:

Yang, X., Ram, N., Robinson, T., & Reeves, B. (2019). Using screenshots to predict task switching on smartphones. ACM Conference on Human Factors in Computing Systems (CHI ’19 Late Breaking Work). Glasgow, UK. May 2019. DOI: 10.1145/3290607.3313089 (pdf)

Ram, N., Yang, X., Cho, M.J., Brinberg, M., Muirhead, F., Reeves, B., & Robinson, T. (2019). Screenomics: A new approach for observing and studying individuals’ digital lives. Journal of Adolescent Research, 1-35, DOI: 10.1177/0743558419883362

Reeves, B., Ram, N., Robinson, T., Cummings, J., Giles, L., Pan, J., Chiatti, Cho, M., Roehrick, K., Yang, X., Gagneja, A., Brinberg, M., Muise, D., Lu, Y., Fitzgerald, A., & Yeykelis, L (2019). Screenomics: a framework to capture and analyze personal life experiences and the ways that technology shapes them. Human-Computer Interaction. DOI: 10.1080/07370024.2019.1578652