Enhancing Personality Recognition by Comparing the Predictive Power of Traits, Facets, and Nuances

TL;DR

This work investigates whether predicting at finer hierarchical levels of the Big Five—facets and nuances—improves self-reported personality recognition from audiovisual dyadic data. Using a transformer-based, interlocutor-aware model on the UDIVA v0.5 dataset, the authors show that nuance-level predictions yield the strongest and most consistent improvements across tasks, outperforming facet- and trait-level models and reducing error metrics substantially. The study demonstrates that fine-grained personality signals capture context-dependent expressions better than aggregate trait scores, with implications for generalization and robustness in personality computing. It also discusses ethical considerations and outlines future work to extend this approach to additional datasets and bias analyses.

Abstract

Personality is a complex, hierarchical construct typically assessed through item-level questionnaires aggregated into broad trait scores. Personality recognition models aim to infer personality traits from different sources of behavioral data. However, reliance on broad trait scores as ground truth, combined with limited training data, poses challenges for generalization, as similar trait scores can manifest through diverse, context dependent behaviors. In this work, we explore the predictive impact of the more granular hierarchical levels of the Big-Five Personality Model, facets and nuances, to enhance personality recognition from audiovisual interaction data. Using the UDIVA v0.5 dataset, we trained a transformer-based model including cross-modal (audiovisual) and cross-subject (dyad-aware) attention mechanisms. Results show that nuance-level models consistently outperform facet and trait-level models, reducing mean squared error by up to 74% across interaction scenarios.

Figures (1)

• Figure 1: Self-reported personality regression pipeline based on dodd2023framework.