EmoPerso: Enhancing Personality Detection with Self-Supervised Emotion-Aware Modelling

2025-09-02

AI Summary

Current personality detection methods rely heavily on large-scale labeled datasets and typically model emotion and personality in isolation, neglecting their intrinsic interdependence. To address these limitations, we propose EmoPerso, an emotion-aware self-supervised framework. First, it employs generative data augmentation to alleviate label scarcity. Second, it introduces a cross-attention mechanism to capture fine-grained interactions between emotion and personality representations. Third, it jointly optimizes emotion pseudo-label generation and personality prediction via multi-task learning, further enhanced by a self-teaching strategy that iteratively refines model generalization. Extensive experiments on two benchmark datasets demonstrate that EmoPerso consistently outperforms existing state-of-the-art methods, validating the effectiveness and robustness of emotion-guided personality detection.

Technology Category

• Humans and AI: Emotional Intelligence
• Machine Learning: Multimodal Learning
• Natural Language Processing: Language Grounding & Multi-modal NLP

Application Category

• User Modeling, Personalization and Recommendation: User privacy protection in personalized systems
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Search and Retrieval-Augmented AI: Personalized, context-aware and across-device search

Abstract

Personality detection from text is commonly performed by analysing users' social media posts. However, existing methods heavily rely on large-scale annotated datasets, making it challenging to obtain high-quality personality labels. Moreover, most studies treat emotion and personality as independent variables, overlooking their interactions. In this paper, we propose a novel self-supervised framework, EmoPerso, which improves personality detection through emotion-aware modelling. EmoPerso first leverages generative mechanisms for synthetic data augmentation and rich representation learning. It then extracts pseudo-labeled emotion features and jointly optimizes them with personality prediction via multi-task learning. A cross-attention module is employed to capture fine-grained interactions between personality traits and the inferred emotional representations. To further refine relational reasoning, EmoPerso adopts a self-taught strategy to enhance the model's reasoning capabilities iteratively. Extensive experiments on two benchmark datasets demonstrate that EmoPerso surpasses state-of-the-art models. The source code is available at https://github.com/slz0925/EmoPerso.

Problem

• Detecting personality from text without large annotated datasets
• Modeling interactions between emotion and personality traits
• Improving personality prediction through self-supervised emotion-aware learning

Innovation

• Self-supervised emotion-aware modeling for personality detection
• Generative synthetic data augmentation and representation learning
• Cross-attention module capturing personality-emotion interactions