đź’ˇ About Me
My research interests mainly focus on  Agentic AI and  Multimodal Affective Intelligence. Welcome to collaborate on related projects.
I am honored to have served as the team leader and won 1st Place in the DES track of the ACM MM 2025 MER (Multimodal Emotion Recognition) Challenge. Additionally, my first-author paper has been accepted by the ACM MM Main Conference Grand Challenge Track.
In 2024, I was recognized as the Guangdong Provincial Person of the Year (one of only 10 winners) and selected as a student representative of the People’s Daily National Scholarship (one of only 4 winners from Guangdong).
| Contact me at: yueshenghuang@stu.gpnu.edu.cn | English / ä¸ć–‡ |
🔥 News
- 2026.01:  🚀 I released Awesome Affective Computing — A curated list of Affective Computing & Emotion AI: Papers, datasets, and toolkits for Multimodal Emotion Recognition, Emotional Reasoning, Multimodal Sentiment Analysis, and Empathetic LLMs/MLLMs. Awesome Affective Computing
- 2025.08:  🏆 I won 1st Place in the ACM MM 2025 MER Challenge (DES Track) as team leader!
- 2025.08: Â đź“„ My first-author paper was accepted by the ACM MM 2025 Main Conference Grand Challenge Track!
- 2024.12:   🎉 I was selected as the 2023 Guangdong Provincial Person of the Year (one of 10 winners), the youngest winner that year.
- 2024.05:   🎉 I was featured in the People’s Daily as a representative of 100 undergraduate national scholarship winners, only 4 of whom were from Guangdong Province.
- 2023.12:   🎉 I was awarded the National Scholarship.
đź“ť Publications
ACM MM 2025
Affective-CoT: Decomposing Multimodal Emotion Reasoning through a Hierarchical Cognitive Workflow
Yuesheng Huang, Jinming Liu, Jiajia Chen, Yihang Lin, Yanmei Chen, Jianwei Dong
BibTeX
@inproceedings{huang2025affective,
title={Affective-CoT: Decomposing Multimodal Emotion Reasoning through a Hierarchical Cognitive Workflow},
author={Huang, Yuesheng and Liu, Jinming and Chen, Jiajia and Lin, Yihang and Chen, Yanmei and Dong, Jianwei},
booktitle={Proceedings of the 33rd ACM International Conference on Multimedia},
pages={13848--13855},
year={2025}
}
title={Affective-CoT: Decomposing Multimodal Emotion Reasoning through a Hierarchical Cognitive Workflow},
author={Huang, Yuesheng and Liu, Jinming and Chen, Jiajia and Lin, Yihang and Chen, Yanmei and Dong, Jianwei},
booktitle={Proceedings of the 33rd ACM International Conference on Multimedia},
pages={13848--13855},
year={2025}
}
MRAC@ACM MM
DARE to Disagree: A Multi-Agent Adversarial Debate Framework for Open-Vocabulary Multimodal Emotion Recognition
Yuesheng Huang, Meiqi Feng, Zhenming He, Yueyuan Peng, Jiawen Li
BibTeX
@inproceedings{huang2025dare,
title={DARE to Disagree: A Multi-Agent Adversarial Debate Framework for Open-Vocabulary Multimodal Emotion Recognition},
author={Huang, Yuesheng and Feng, Meiqi and He, Zhenming and Peng, Yueyuan and Li, Jiawen},
booktitle={Proceedings of the 3rd International Workshop on Multimodal and Responsible Affective Computing},
pages={41--50},
year={2025}
}
title={DARE to Disagree: A Multi-Agent Adversarial Debate Framework for Open-Vocabulary Multimodal Emotion Recognition},
author={Huang, Yuesheng and Feng, Meiqi and He, Zhenming and Peng, Yueyuan and Li, Jiawen},
booktitle={Proceedings of the 3rd International Workshop on Multimodal and Responsible Affective Computing},
pages={41--50},
year={2025}
}
ICML NewInML
Can Generated Images Serve as a Viable Modality for Text-Centric Multimodal Learning?
Yuesheng Huang, Peng Zhang, Riliang Liu, Jiaqi Liang
BibTeX
@misc{huang2025generatedimagesserveviable,
title={Can Generated Images Serve as a Viable Modality for Text-Centric Multimodal Learning?},
author={Yuesheng Huang and Peng Zhang and Riliang Liu and Jiaqi Liang},
year={2025},
eprint={2506.17623},
archivePrefix={arXiv},
primaryClass={cs.MM},
url={https://arxiv.org/abs/2506.17623},
}
title={Can Generated Images Serve as a Viable Modality for Text-Centric Multimodal Learning?},
author={Yuesheng Huang and Peng Zhang and Riliang Liu and Jiaqi Liang},
year={2025},
eprint={2506.17623},
archivePrefix={arXiv},
primaryClass={cs.MM},
url={https://arxiv.org/abs/2506.17623},
}
CMC 2024
Sentiment Analysis Using E-Commerce Review Keyword-Generated Image with a Hybrid Machine Learning-Based Model
Jiawen Li, Yuesheng Huang, Yayi Lu, Leijun Wang*, Yongqi Ren and Rongjun Chen
GB/T
Li J, Huang Y, Lu Y, et al. Sentiment Analysis Using E-Commerce Review Keyword-Generated Image with a Hybrid Machine Learning-Based Model[J]. Computers, Materials & Continua, 2024, 80(1).
MLA
Li, Jiawen, et al. "Sentiment Analysis Using E-Commerce Review Keyword-Generated Image with a Hybrid Machine Learning-Based Model." Computers, Materials & Continua 80.1 (2024).
APA
Li, J., Huang, Y., Lu, Y., Wang, L., Ren, Y., & Chen, R. (2024). Sentiment Analysis Using E-Commerce Review Keyword-Generated Image with a Hybrid Machine Learning-Based Model. Computers, Materials & Continua, 80(1).
Engineering Letters
An Improved Hybrid CNN-LSTM-Attention Model with Kepler Optimization Algorithm for Wind Speed Prediction
Yuesheng Huang, Jiawen Li, Yushan Li, Routing Lin, Jingru Wu, Leijun Wang, and Rongjun Chen
GB/T
Huang Y, Li J, Li Y, et al. An Improved Hybrid CNN-LSTM-Attention Model with Kepler Optimization Algorithm for Wind Speed Prediction[J]. Engineering Letters, 2024, 32(10).
MLA
Huang, Yuesheng, et al. "An Improved Hybrid CNN-LSTM-Attention Model with Kepler Optimization Algorithm for Wind Speed Prediction." Engineering Letters 32.10 (2024).
APA
Huang, Y., Li, J., Li Y., Lin, R., Wu, J., Wang, L., & Chen, R. (2024). An Improved Hybrid CNN-LSTM-Attention Model with Kepler Optimization Algorithm for Wind Speed Prediction. Engineering Letters, 32(10).
🏆 Competition Awards
- 2025.08 Champion (1st Place) in the ACM MM 2025 MER Challenge (DES Track) as team leader.
- 2025.02 First Prize in Information Technology, Healthcare, and Modern Service Tracks at the China University Student Technology Innovation and Entrepreneurship Competition.
- 2024.06 Silver Award in the Challenge Cup Entrepreneurship Plan Competition (Guangdong), Guangdong Provincial Department of Education.
- 2024.05 Finalist Award in the COMAP Mathematical Contest in Modeling (MCM/ICM), Problem E (Top 2% worldwide).
- 2023.11 First Prize in the Guangdong Division of the China Undergraduate Mathematical Contest in Modeling, Guangdong Provincial Department of Education.
- 2023.08 First Prize in the China College Student Computer Design Competition (Guangdong Division), Guangdong Provincial Department of Education.
- 2023.08 Outstanding Award in the International University Mathematical Contest in Modeling.
- 2023.07 First Prize in the National Undergraduate Electrical Mathematical Modeling Competition, China Electrotechnical Society.
- 2021.12 Silver Medal in the Kaggle Lux AI Competition.
🎓 Educations
- 2021.09 - 2025.06, Bachelor of Engineering in Internet of Things Engineering(ESI TOP 1%), School of Computer Science, Guangdong Polytechnic Normal University, Excellent Graduate.(GPA:92.1/100, Rank:1/111) (Highest score in the cohort)
Design of an Intelligent Student Emotion Analysis and Monitoring System Empowered by Multimodal Data and Large ModelsABSTRACT
With the in-depth development of artificial intelligence and deep learning technologies, the application potential of multimodal sentiment analysis in the educational field is increasingly evident. Traditional unimodal emotion recognition methods have limitations in capturing students' complex emotional states, while multimodal analysis significantly improves the accuracy of emotion recognition by integrating facial expressions, speech information, and physiological signals. Under the current background of educational informatization, there is an urgent demand for student mental health monitoring. However, existing methods face challenges such as poor timeliness, strong subjectivity, and difficulties in scaling, which limit their widespread adoption in campus environments.
To address these challenges, this paper proposes and implements a student emotion intelligent analysis and monitoring system based on ESP32 and ESP32S3 hardware platforms, combined with lightweight multimodal fusion algorithms and large language models. The system aims to utilize low-cost, highly integrated embedded technology to fuse multi-source data including facial, speech, and heart rate information, providing educators, parents, and students with real-time, accurate, and convenient emotion monitoring and support tools.
At the hardware level, a distributed dual-mainboard architecture using ESP32 and ESP32S3 is adopted. The ESP32 mainboard integrates ESP32CAM and heart rate sensors to implement facial expression recognition, physiological data collection, and basic feedback. The ESP32S3 mainboard integrates a digital microphone, audio amplifier, display screen, etc., to achieve intelligent dialogue functions based on Baidu's ERNIE Bot API. At the software level, a Node.js-based server is constructed, SQLite is used for data storage, and multi-role web application interfaces for teachers, students, and parents are developed. On the algorithmic level, the system implements facial emotion recognition based on Deepface, speech emotion analysis using the ERNIE Bot API, designs a dynamic weight decision-level multimodal fusion algorithm, and introduces a data-volume-based multi-model emotion trend prediction method. Additionally, prompt optimization is employed to enhance the performance of large language models in emotional support dialogue tasks.
Finally, the system hardware platform was successfully built and debugged, with comprehensive functional testing and verification conducted on the software system, including white-box testing and black-box testing. The test results demonstrate stable operation of all system modules, compliance with design requirements, and effective integration of multimodal data for student emotion state analysis and monitoring, validating the feasibility and effectiveness of the design.
Keywords: Multimodal sentiment analysis; Student emotion monitoring; ESP32; Large language models; Data fusion - 2018.09 - 2021.06, Ordinary high school, Shaoguan City Wengyuan middle School
đź“– Research topics
- 2023.05-2024.05, “Research and implementation of MIMO system detection algorithm based on Gaussian tree”, Chinese college students Innovation and Entrepreneurship plan project, Huang Yuesheng as host. (Project completed)
- 2024.01-2026-01, “Research on fine-grained sentiment analysis of multi-modal data fusion based on deep learning”, Guangdong Provincial Science and Technology Innovation Fund, 45,000CNY, Huang Yuesheng as host. (Project completed)
- 2024.05-2025.05, “Neurodetective: An interpretable multimodal contrastive learning Framework for the diagnosis of neurodegenerative diseases”, Chinese college students Innovation and Entrepreneurship plan project, Huang Yuesheng as host. (Project completed)
- 2024.05-2025.05, “Aquaponics, Ecological co-prosperity: A general agricultural visual large model for digital aquaponics fish pond system called DASAM”, Chinese college students Innovation and Entrepreneurship plan project, Participant.(Project completed)
©️ Patents and Copyrights
- 2025, “Multimodal Disease Diagnosis Software Based on Diffusion-Model Image Synthesis V1.0”, Chinese software copyright, 2025SR1592713, First contributor
- 2024, “Flask based medical image segmentation platform V1.0”, Chinese software copyright, 2024SR0877362