Author(s)

Lin Gan ¹, Zhengpeng Zhang ¹

Affiliation(s)

¹ School of Information and Intelligent Engineering, University of Sanya, Sanya, Hainan, 572100, China

Corresponding Author

Lin Gan

ABSTRACT

The development of multi-modal large models provides a robust representation foundation for sentiment analysis. However, current research primarily focuses on static classification tasks, neglecting the dynamic evolutionary nature of emotions. This paper centers around the core concept of the "emotion causal chain," systematically reviewing the research status of multi-modal feature fusion and counterfactual learning in the field of sentiment analysis. Based on defining the theoretical connotation of the emotion causal chain, it critically compares early fusion, attention mechanisms, graph neural networks, and large model fusion paradigms from a technological evolution perspective, pointing out the fundamental limitations of existing methods in terms of interaction range, spurious correlation control, and interpretability. Furthermore, it focuses on discussing the theoretical foundation and application paths of counterfactual learning, elucidating its methodological advantages in modal decoupling, temporal intervention, and path identification. It also systematically summarizes the implementation framework based on generative models, contrastive learning, causal attention, and large model integration. The research suggests that counterfactual learning enables models to go beyond statistical associations and touch upon the causal mechanisms of emotion evolution, enhancing interpretability and robustness while providing computational tools for analyzing cross-modal emotion transmission paths. Finally, it looks forward to future directions, emphasizing that building a multi-modal fusion model with causal reasoning ability is the key to achieving interpretable and strongly generalizable emotional intelligence.

KEYWORDS

Multi-modal Fusion; Emotion Causal Chain; Counterfactual Learning; Causal Inference; Sentiment Analysis

CITE THIS PAPER

Lin Gan, Zhengpeng Zhang. Emotion Causal Chain Analysis Method Based on Multi-Modal Feature Fusion. Journal of Artificial Intelligence Practice (2026). Vol. 9, No. 1, 99-107. DOI: http://dx.doi.org/10.23977/jaip.2026.090112.

REFERENCES

[1] Jie Liliang, Zou Yangmeng, Li Zhengxiu, et al. Method for Emotion Conversion Recognition Based on Cross-modal Feature Fusion and Global Perception [J]. Journal of Biomedical Engineering, 2025, 42(05): 977-986.
[2] Zhang Zhiwen, Yu Nai Gong, Bian Yan, et al. Research on Emotion Recognition Based on Multi-modal Physiological Signal Feature Fusion [J]. Journal of Biomedical Engineering, 2025, 42(01): 17-23.
[3] Zhang Yiqing. Research on Emotional Understanding Method Based on Multi-modal Feature Fusion [D]. Beijing University of Posts and Telecommunications, 2025.
[4] Gao Jingjing. Research on Multi-modal Emotion Recognition Based on Multi-dimensional Features and CRNN [D]. Qufu Normal University, 2025.
[5] Wang Ke. Research on Emotion Recognition Method Based on Multi-modal Fusion [D]. Jilin University, 2025.
[6] Zheng Peng. Research on Audio-Video Emotional Analysis Based on Multi-modal Feature Fusion [D]. Jiangxi Normal University, 2025.
[7] Fu Junsong. Research on Emotional Analysis Based on Multi-modal Feature Fusion [D]. Chongqing University of Technology, 2025.
[8] Yang Liang. Research on Emotional Analysis Based on Multi-modal Feature Fusion [D]. Guangxi Normal University, 2025.

Sponsors, Associates, and Links

---

• Power Systems Computation
  
  
• Internet of Things (IoT) and Engineering Applications
  
  
• Computing, Performance and Communication Systems
  
  
• Advances in Computer, Signals and Systems
  
  
• Journal of Network Computing and Applications
  
  
• Journal of Web Systems and Applications
  
  
• Journal of Electrotechnology, Electrical Engineering and Management
  
  
• Journal of Wireless Sensors and Sensor Networks
  
  
• Journal of Image Processing Theory and Applications
  
  
• Mobile Computing and Networking
  
  
• Vehicle Power and Propulsion
  
  
• Frontiers in Computer Vision and Pattern Recognition
  
  
• Knowledge Discovery and Data Mining Letters
  
  
• Big Data Analysis and Cloud Computing
  
  
• Electrical Insulation and Dielectrics
  
  
• Crypto and Information Security
  
  
• Journal of Neural Information Processing
  
  
• Collaborative and Social Computing
  
  
• International Journal of Network and Communication Technology
  
  
• File and Storage Technologies
  
  
• Frontiers in Genetic and Evolutionary Computation
  
  
• Optical Network Design and Modeling
  
  
• Journal of Virtual Reality and Artificial Intelligence
  
  
• Natural Language Processing and Speech Recognition
  
  
• Journal of High-Voltage
  
  
• Programming Languages and Operating Systems
  
  
• Visual Communications and Image Processing
  
  
• Journal of Systems Analysis and Integration
  
  
• Knowledge Representation and Automated Reasoning
  
  
• Review of Information Display Techniques
  
  
• Data and Knowledge Engineering
  
  
• Journal of Database Systems
  
  
• Journal of Cluster and Grid Computing
  
  
• Cloud and Service-Oriented Computing
  
  
• Journal of Networking, Architecture and Storage
  
  
• Journal of Software Engineering and Metrics
  
  
• Visualization Techniques
  
  
• Journal of Parallel and Distributed Processing
  
  
• Journal of Modeling, Analysis and Simulation
  
  
• Journal of Privacy, Trust and Security
  
  
• Journal of Cognitive Informatics and Cognitive Computing
  
  
• Lecture Notes on Wireless Networks and Communications
  
  
• International Journal of Computer and Communications Security
  
  
• Journal of Multimedia Techniques
  
  
• Automation and Machine Learning
  
  
• Computational Linguistics Letters
  
  
• Journal of Computer Architecture and Design
  
  
• Journal of Ubiquitous and Future Networks