The scientific research poster has been a mainstay for decades. For USC Viterbi magazine, we took a recent poster from USC Viterbi Assistant Professor Yan Liu’s Melady Lab in the Department of Computer Science, and re-imagined the ideas in various art forms.


By mining social media data, can Big Data analytic algorithms detect anomalous behaviors, such as cyber-bullying, terrorist planning or fraud dissemination? Liu’s team developed GLAD (Group Latent Anomaly Detection), a statistical machine-learning model that automatically finds anomalies in groups of users, as opposed to just individuals. To achieve this goal, Liu’s team applies this model to both point-wise data (individual e-mails, posts) and pair-wise data (exchange of e-mails) to assign a ranking (0 to 1) of groups in terms of anomalousness by comparing with other similar groups.

— Adam Smith



As “Where’s Waldo?” Book

Illustration: Douglas Holgate


As “Divergent” Poster Art



As Esquire Cover

Illustration: Michelle Henry


As “Peanuts” Comic Strip

Illustration: Katherine Duffy


As Chinese Watercolor