R1-Omni: Explainable Omni-Multimodal Emotion Recognition with Reinforcing Learning

2025-03-07

AI Summary

This paper addresses three key challenges in vision-audio multimodal emotion recognition: weak logical reasoning, low accuracy, and poor out-of-distribution (OOD) generalization. To this end, we propose an Omni-Modality Large Language Model (Omni-MLLM) optimization framework grounded in verifiable-reward reinforcement learning (RLVR). This work is the first to introduce RLVR into multimodal emotion recognition, jointly enhancing reasoning coherence, classification accuracy, and OOD robustness. We further design a novel modality attribution mechanism that enables interpretable and quantifiable analysis of visual and auditory contributions. Extensive experiments demonstrate that our method significantly outperforms state-of-the-art approaches on both in-distribution and OOD benchmarks, achieving superior accuracy and strong generalization. The framework establishes a new paradigm for multimodal emotion modeling—verifiable through reward signals and attributable via modality-wise contribution analysis.

Technology Category

• Computer Vision: Multi-modal Vision
• Machine Learning: Multimodal Learning
• Natural Language Processing: Language Grounding & Multi-modal NLP

Application Category

• Search and Retrieval-Augmented AI: Retrieval-Augmented Generation (RAG) and multi-modal RAG
• User Modeling, Personalization and Recommendation: Large Language Models (LLM) for user modeling and recommendation
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs

Abstract

In this work, we present the first application of Reinforcement Learning with Verifiable Reward (RLVR) to an Omni-multimodal large language model in the context of emotion recognition, a task where both visual and audio modalities play crucial roles. We leverage RLVR to optimize the Omni model, significantly enhancing its performance in three key aspects: reasoning capability, emotion recognition accuracy, and generalization ability. The introduction of RLVR not only improves the model's overall performance on in-distribution data but also demonstrates superior robustness when evaluated on out-of-distribution datasets. More importantly, the improved reasoning capability enables clear analysis of the contributions of different modalities, particularly visual and audio information, in the emotion recognition process. This provides valuable insights into the optimization of multimodal large language models.

Problem

• Enhance emotion recognition using Reinforcement Learning with Verifiable Reward.
• Improve reasoning, accuracy, and generalization in multimodal emotion recognition.
• Analyze contributions of visual and audio modalities in emotion recognition.

Innovation

• Reinforcement Learning with Verifiable Reward (RLVR)
• Omni-multimodal large language model
• Enhanced reasoning and emotion recognition accuracy