On October 29, Dr. Andrew Schwartz presented “AI Language Analyses for Mental Health: Digital Words as a Powerful Marker of Well-Being” at the University Libraries.
Traditionally, surveys, interviews, and observations are commonly used for assessments. For the first time, Dr. Schwartz combines language and machine learning as modern survey to analyze and look for patterns of mental health issues such as depressing through social media, for instance, Facebook and Twitter. This AI-based language analysis provides a good predictive mental health evaluation when integrating into other data analysis.
Dr. Schwartz also found that Twitter is a good tool to capture an abundance of information such as predicting heart disease in the community level. Using age as a criteria, Dr. Schwartz analyzed words used in social media by different age groups and discovered that each group conversed within the same languages that are appropriate to their ages. Inspired by the National Child Development Study, a longitudinal project started in 1958 by University College London (UCL), Dr. Schwartz used the available written data by participants to conduct his own innovative research.
The lecture was well attended by various audiences including clinical students and faculty. It generated a rich discussion about language-based assessment using AI.
Latest posts by Clara Tran (see all)
- Dr. Thomas Woodson on “Should research have societal impact? Re-evaluating broader impacts with the Inclusion-Immediacy Criterion” - October 12, 2020
- Dr. Carlos Simmerling on “Using computer simulations to model the SARS-CoV-2 spike glycoprotein and block COVID-19 infection” - October 6, 2020
- “Should research have societal impact? Re-evaluating broader impacts with the Inclusion-Immediacy Criterion” with Dr. Thomas Woodson - September 29, 2020