Recent Publications, Patents & Software

Here are some of my recently published academic papers, covering topics such as multimodal learning, large language models, vision-language-action (VLA) agents, embodied AI, and affective computing. With 500+ citations to date, my full publication list (20+) is available at my Google Scholar page.

Papers as first or corresponding author:

• MIRTH: Mutual-Information Reasoning with Temporal Hubs for Vision-Language-Action Agents. The 64th Annual Meeting of the Association for Computational Linguistics (ACL), 2026.
• One Framework to Rule Them All: Unifying Multimodal Tasks with LLM Neural-Tuning. Pattern Recognition, 2025. (IF: 8.5)
• Multimodal Sentiment Analysis with Mutual Information-based Disentangled Representation Learning. IEEE Transactions on Affective Computing, Vol.16(3), pp.1606-1617, 2025. (IF: 13.9)
• Modality-invariant Temporal Representation Learning for Multimodal Sentiment Classification. Information Fusion, Vol.91, pp.504-514, 2023. (IF: 18.1)
• Tensorformer: A Tensor-Based Multimodal Transformer for Multimodal Sentiment Analysis and Depression Detection. IEEE Transactions on Affective Computing, Vol.14(4), pp.2776-2786, 2023. (IF: 13.9)
• Multi-Modal Adaptive Fusion Transformer Network for the Estimation of Depression Level. Sensors, Vol.21(14), pp.4764, 2021.
• CubeMLP: An MLP-Based Model for Multimodal Sentiment Analysis and Depression Estimation. ACM Multimedia, pp.3722-3729, 2022. (Cited 140+)

Preprints in submission (first author):

• Multimodal Infusion Tuning for Large Models. arXiv:2403.05060, 2024.
• Robust Latent Representation Tuning for Image-text Classification. arXiv:2406.06048, 2024.
• Modality-invariant and Specific Prompting for Multimodal Human Perception Understanding. arXiv:2311.10791, 2023.

Papers as co-first, second or third author:

• EPIC: Efficient Prompt Interaction for Text-Image Classification. IEEE ICME, 2025. (Co-First Author)
• CG-DMER: Hybrid Contrastive-Generative Framework for Disentangled Multimodal ECG Representation Learning. IEEE ICASSP, 2026.
• Improving scDiffusion with Sparsity-Biased Classifier-Free Guidance. IEEE Engineering in Medicine and Biology Society, 2026.
• Dynamic Summary Generation for Interpretable Multimodal Depression Detection. IEEE ICASSP, 2026.
• IRLSG: Invariant Representation Learning for Single-Domain Generalization in Medical Image Segmentation. IEEE ICASSP, 2024.
• LGA: A Language Guide Adapter for Advancing the SAM Model's Capabilities in Medical Image Segmentation. Springer MICCAI, pp.610-620, 2024.
• Enhanced Multimodal Depression Detection With Emotion Prompts. IEEE ICASSP, 2025.
• DepressionLLM: Emotion- and Causality-aware Depression Detection with Foundation Models. Displays, Vol.92, pp.103304, 2025.
• MCKD: Mutually Collaborative Knowledge Distillation for Federated Domain Adaptation and Generalization. IEEE ICASSP, 2023.
• CoSTHR: A Heart Rate Estimating Network with Adaptive Color Space Transformation. IEEE Transactions on Instrumentation and Measurement, Vol.71, pp.1-10, 2022. (IF: 5.6)

Recent Patents:

• Engineering Progress Determining Method and Device Based on Multi-Mode Time Sequence Information Fusion. App. No: US20250005475A1. Country: US. Publication Date: 2023-07. Third Inventor
• Engineering Progress Determining Method and Device Based on Multi-Mode Time Sequence Information Fusion. Grant No: CN116502882B. Country: CN. Grant Date: 2023-10. Third Inventor
• Rheumatoid Arthritis Activity Grading Device Based on Multimodal Data. Grant No: CN116797572B. Country: CN. Grant Date: 2025-09. Fourth Inventor
• A Single-Domain Generalization Method for Medical Image Segmentation. Grant No: CN116596832B. Country: CN. Grant Date: 2025-07. Fifth Inventor.

Software Copyrights:

• MIRTH Robot Control System V1.0. Registration No.: 2026SR0601586. Country: China. Authorization Date: May 2026.

For 30+ independent projects, please visit my Github Page.
From 2020 to 2021, I was responsible for publishing TensorFlow tutorials on IMOOC.

Skills

This section outlines my core academic and engineering skills, including research design, algorithm development, large-scale model training, and practical system implementation, with a strong focus on AI, large language models (LLMs), multimodal learning, and embodied AI.

• Academic-Specific: Scholarly Writing, Publication, Peer Review, Conference Presentation, Grant Proposal
• AI Research: Algorithm Development, Model Training & Finetuning, Data Processing, Evaluation
• LLM-Specific: Model Customization, Knowledge Integration, Multimodal Tuning, Scalability and Efficiency
• Embodied AI: Action Grounding, Platform Building, Hardware Employment
• Multimodal Research: Framework Design, Task Adaptation, Multimodal System Deployment
• Software Engineering: Feasibility and Requirements Analysis, System and Detailed Design, Implementation and Software Maintenance, etc
• Programming Languages: Python, PyTorch, Numpy, TensorFlow, Java, C++, C, HTML, GO, etc
• Full-stack Development: FrontEnd Programming, BackEnd Design and Implementation, Database Systems
• Language: Mandarin, English, Japanese