What is it about?

We use language models to improve our understanding of human personality. Our approach accurately predicts how people would respond to new personality questions. It can also identify psychological traits in survey questions and group them together based on their language alone. This innovation offers a cost-effective way to measure psychological traits.

Featured Image

Why is it important?

Our research leverages machine learning and advanced language processing to achieve human-level accuracy in predicting ratings for new items. This is relevant in two ways. First, it allows to develop cost-effective and scalable methods for psychological measurement increasing the amount of information extracted from data collection, particularly useful for understudied traits. Second, our approach allows for analytical examinations of personality and broader psychological constructs via their quantitative semantic representations, offering a potent tool for advancing psychological theories.


Working with my collaborators on this research has been a delightful experience. Our work in this area has led to a series of interesting projects in scale development and the study of social inferences in humans and (large) language models. It's been a rewarding journey that promises to enhance our understanding of these important aspects of human psychology and artificial intelligence.

Suhaib Abdurahman
University of Southern California

Read the Original

This page is a summary of: A deep learning approach to personality assessment: Generalizing across items and expanding the reach of survey-based research., Journal of Personality and Social Psychology, September 2023, American Psychological Association (APA),
DOI: 10.1037/pspp0000480.
You can read the full text:




The following have contributed to this page