FATRER: Full-Attention Topic Regularizer for Accurate and Robust Conversational Emotion Recognition

TL;DR

This paper introduces FATRER, a novel emotion recognition model that combines full-attention topic regularization with joint topic modeling to improve accuracy and robustness against adversarial attacks in conversational settings.

Contribution

The paper proposes a full-attention topic regularizer and a joint topic modeling strategy that enhance robustness and accuracy in conversational emotion recognition.

Findings

Outperforms state-of-the-art models in accuracy.

Demonstrates improved robustness under adversarial attacks.

Effective in maintaining emotion recognition performance with corrupted local context.

Abstract

This paper concentrates on the understanding of interlocutors' emotions evoked in conversational utterances. Previous studies in this literature mainly focus on more accurate emotional predictions, while ignoring model robustness when the local context is corrupted by adversarial attacks. To maintain robustness while ensuring accuracy, we propose an emotion recognizer augmented by a full-attention topic regularizer, which enables an emotion-related global view when modeling the local context in a conversation. A joint topic modeling strategy is introduced to implement regularization from both representation and loss perspectives. To avoid over-regularization, we drop the constraints on prior distributions that exist in traditional topic modeling and perform probabilistic approximations based entirely on attention alignment. Experiments show that our models obtain more favorable results than state-of-the-art models, and gain convincing robustness under three types of adversarial attacks.