AffectGPT-RL: Revealing Roles of Reinforcement Learning in Open-Vocabulary Emotion Recognition

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Lian, Zheng, Zhang, Fan, Chen, Lan, Zhang, Yazhou, Liu, Rui, Wu, Jinyang, Chen, Haoyu, Li, Xiaobai, Peng, Xiaojiang, He, Bin, Tao, Jianhua
Format: Preprint
Published: 2026
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866914539557093376
author Lian, Zheng
Zhang, Fan
Chen, Lan
Zhang, Yazhou
Liu, Rui
Wu, Jinyang
Chen, Haoyu
Li, Xiaobai
Peng, Xiaojiang
He, Bin
Tao, Jianhua
author_facet Lian, Zheng
Zhang, Fan
Chen, Lan
Zhang, Yazhou
Liu, Rui
Wu, Jinyang
Chen, Haoyu
Li, Xiaobai
Peng, Xiaojiang
He, Bin
Tao, Jianhua
contents Open-Vocabulary Multimodal Emotion Recognition (OV-MER) aims to predict emotions without being constrained by predefined label spaces, thereby enabling fine-grained emotion understanding. Unlike traditional discriminative methods, OV-MER leverages generative models to capture the full spectrum of emotions and employs emotion wheels (EWs) for metric calculation. Previous approaches primarily rely on token-level loss during training. However, this objective is misaligned with the metrics used in OV-MER, and these metrics cannot be directly optimized via gradient backpropagation. To address this limitation, we turn our attention to reinforcement learning, as this strategy can optimize non-differentiable objectives. We term this framework AffectGPT-RL. Furthermore, we conduct extensive experiments to elucidate the role of reinforcement learning in this task, revealing the necessity of the reasoning process, the impact of different rewards, and the generalizability to other emotion tasks such as sentiment analysis and basic emotion recognition. Experimental results demonstrate that AffectGPT-RL yields significant performance improvements on OV-MER. Beyond this task, we also achieve remarkable performance gains on basic emotion recognition, attaining state-of-the-art results on MER-UniBench. To the best of our knowledge, this is the pioneering work exploring the role of reinforcement learning in OV-MER, providing valuable guidance for subsequent researchers. Our code is provided in the supplementary material and will be released to facilitate future research.
format Preprint
id arxiv_https___arxiv_org_abs_2605_06126
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle AffectGPT-RL: Revealing Roles of Reinforcement Learning in Open-Vocabulary Emotion Recognition
Lian, Zheng
Zhang, Fan
Chen, Lan
Zhang, Yazhou
Liu, Rui
Wu, Jinyang
Chen, Haoyu
Li, Xiaobai
Peng, Xiaojiang
He, Bin
Tao, Jianhua
Human-Computer Interaction
Open-Vocabulary Multimodal Emotion Recognition (OV-MER) aims to predict emotions without being constrained by predefined label spaces, thereby enabling fine-grained emotion understanding. Unlike traditional discriminative methods, OV-MER leverages generative models to capture the full spectrum of emotions and employs emotion wheels (EWs) for metric calculation. Previous approaches primarily rely on token-level loss during training. However, this objective is misaligned with the metrics used in OV-MER, and these metrics cannot be directly optimized via gradient backpropagation. To address this limitation, we turn our attention to reinforcement learning, as this strategy can optimize non-differentiable objectives. We term this framework AffectGPT-RL. Furthermore, we conduct extensive experiments to elucidate the role of reinforcement learning in this task, revealing the necessity of the reasoning process, the impact of different rewards, and the generalizability to other emotion tasks such as sentiment analysis and basic emotion recognition. Experimental results demonstrate that AffectGPT-RL yields significant performance improvements on OV-MER. Beyond this task, we also achieve remarkable performance gains on basic emotion recognition, attaining state-of-the-art results on MER-UniBench. To the best of our knowledge, this is the pioneering work exploring the role of reinforcement learning in OV-MER, providing valuable guidance for subsequent researchers. Our code is provided in the supplementary material and will be released to facilitate future research.
title AffectGPT-RL: Revealing Roles of Reinforcement Learning in Open-Vocabulary Emotion Recognition
topic Human-Computer Interaction
url https://arxiv.org/abs/2605.06126