Research profile
AI researcher focused on medical multimodal systems.
Published in Artificial Intelligence in Medicine, with research experience across A*STAR collaboration, SUTD, medical image analysis, and LLM systems.
Medical AI, multimodal learning, LLM reliability
Lihan Zuo
I earned my B.Eng. in Artificial Intelligence from Southwest Jiaotong University (SWJTU) in 2023, previously worked as a Visiting Researcher at ISTD, SUTD, and now work as an Algorithm Engineer at LexinFintech in Shenzhen.
Research Focus
Robust multimodal intelligence.
Three research tracks: medical multimodal learning, reliable LLM workflows, and memory-aware language models.
01
/ 03
01 / Medical AI
Adaptive multimodal fusion for skin lesion classification
Published work in Artificial Intelligence in Medicine on combining clinical images, dermoscopy images, and metadata for multi-label skin lesion classification.
02 / LLM
Hallucination mitigation and reliable LLM workflows
Open-source exploration around full-loop LLM workflows, model behavior analysis, and methods for reducing hallucination in practical systems.
03 / LLM Memory
Memory-aware language models and long-context reasoning
Research interest in how large language models store, retrieve, revise, and use memory across long interactions and knowledge-intensive workflows.
Publications
Peer-reviewed research papers.
A multi-stage multi-modal learning algorithm with adaptive multimodal fusion for improving multi-label skin lesion classification
Peer-reviewed medical AI paper on multimodal skin lesion classification using clinical images, dermoscopy images, and metadata.
The work proposes an uncertainty-based hybrid fusion strategy for multimodal skin cancer diagnosis, combining image-level and metadata-level information for more adaptive classification.
@article{zuo2025multistage,
title = {A multi-stage multi-modal learning algorithm with adaptive multimodal fusion for improving multi-label skin lesion classification},
author = {Zuo, Lihan and Wang, Zizhou and Wang, Yan},
journal = {Artificial Intelligence in Medicine},
volume = {162},
pages = {103091},
year = {2025},
doi = {10.1016/j.artmed.2025.103091}
}
Projects
Research systems as inspectable artifacts.
01Medical AI
CosCatNet adaptive fusion algorithm
Experiment code for the AIIM 2025 paper on multi-stage multimodal skin lesion classification.
02Dementia AI
Multi-Modality Dementia Detection Model
Working on the Multi-Modality Dementia Detection Model with members from SUTD and Icekredit. 🇸🇬
CV
Academic CV, stitched as a living timeline.
A structured academic timeline covering education, research experience, professional work, projects, and publication record.
2025.06 - Now
Algorithm Engineer, LexinFintech Holdings Ltd.
Shenzhen, China
Professional role focused on algorithm engineering, model development, and applied AI systems.
2025
Artificial Intelligence in Medicine publication
Lihan Zuo, Zizhou Wang, Yan Wang
First-author paper: A multi-stage multi-modal learning algorithm with adaptive multimodal fusion for improving multi-label skin lesion classification.
2024.05 - 2025.05
LLM research at SUTD with A*STAR collaboration
Research experience, Singapore
Worked on large language models at SUTD and collaborated with scientists from A*STAR.
2024.05 - 2025.05
Multi-Modality Dementia Detection Model
Project with members from SUTD and IceKredit
Applied multimodal AI project focused on dementia detection with cross-institution collaboration.
2023.09 - 2024.12
Medical image analysis with A*STAR scientists
Research experience, Singapore
Worked on medical image analysis, leading into the multimodal skin lesion classification publication.
2019.09 - 2023.06
B.Eng. in Artificial Intelligence, Southwest Jiaotong University (SWJTU)
SWJTU, China
Undergraduate training in artificial intelligence, with foundations in machine learning, programming, and intelligent systems.
Docs
Operational notes for reproducible research.
Contact