SURE: Synergistic Uncertainty-aware Reasoning for Multimodal Emotion Recognition in Conversations

TL;DR

SURE is a novel framework for multimodal emotion recognition in conversations that enhances robustness and contextual understanding by modeling uncertainty, iterative reasoning, and modal interactions.

Contribution

It introduces a comprehensive uncertainty-aware reasoning framework with three key modules, advancing robustness and accuracy in MERC tasks.

Findings

SURE outperforms existing methods on benchmark datasets.

Uncertainty modeling improves noise robustness in multimodal signals.

Iterative reasoning enhances contextual understanding in conversations.

Abstract

Multimodal emotion recognition in conversations (MERC) requires integrating multimodal signals while being robust to noise and modeling contextual reasoning. Existing approaches often emphasize fusion but overlook uncertainty in noisy features and fine-grained reasoning. We propose SURE (Synergistic Uncertainty-aware REasoning) for MERC, a framework that improves robustness and contextual modeling. SURE consists of three components: an Uncertainty-Aware Mixture-of-Experts module to handle modality-specific noise, an Iterative Reasoning module for multi-turn reasoning over context, and a Transformer Gate module to capture intra- and inter-modal interactions. Experiments on benchmark MERC datasets show that SURE consistently outperforms state-of-the-art methods, demonstrating its effectiveness in robust multimodal reasoning. These results highlight the importance of uncertainty modeling and iterative reasoning in advancing emotion recognition in conversational settings.