Detecting Personality during Group Interactions from Multimodal Behaviors

Description: This project is to detect personality from multimodal behavioural data within small groups. You will use a dataset (as shown in the figure) that recorded videos when 3 or 4 participants had a group discussion (Müller et al. 2018). Participants also reported their personality by completing a Big Five personality questionnaire. Speech, body movement, posture, eye contact and facial expression can be extracted from the video to predict a participant’s personality via machine learning or deep learning.

Supervisor: Guanhua Zhang

Distribution: 20% Literature, 20% Data Preparation, 40% Implementation, 20% Data Analysis

Requirements: Strong programming skills, experience with machine learning/deep learning and data analysis

Literature: Philipp Müller, Michael Xuelin Huang, and Andreas Bulling. 2018. Detecting Low Rapport During Natural Interactions in Small Groups from Non-Verbal Behaviour. Proceedings of the 23rd International Conference on Intelligent User Interfaces (IUI).