About me

My research interests include intelligent agents 🤖 and spatial reasoning 🧭. I aim to explore how agents can perceive, represent, and reason about information from multi-modal data, thereby improving the robustness and reliability of intelligent systems. I believe that grounding agentic intelligence in effective representation and reasoning can provide a strong foundation for decision-making, and further support trustworthy applications in complex real-world scenarios.

I am currently pursuing a Master of Science in Computer Vision (M.Sc.) at Mohamed bin Zayed University of Artificial Intelligence (MBZUAI).

Publications

You can also find a list of my published work on Google Scholar.

Medical Image Analysis

1. Global contrast-masked autoencoders are powerful pathological representation learners
   Hao Quan^*, Xingyu Li^* (co-first author), Weixing Chen, Qun Bai, Mingchen Zou, Ruijie Yang, Tingting Zheng, Ruiqun Qi, Xinghua Gao, Xiaoyu Cui
   Pattern Recognition (PR)(2024) [PDF] [BibTeX] [Source Code]

2. Decoding Expertise from Pathologists’ Diagnostic Processes on Whole Slide Images
   Nature Communications (2025) [PDF] [Source Code]

3. Discriminating Chromophobe RCC from Oncocytoma: A Transformer-based Approach Leveraging the Subtleties of Nuclear Structures within Kidney Tumors
   Jing Yang, Xingyu Li, Hongjiu Ren, Yanmei Zhu, Qimin Wang, Ruiqun Qi, Xiaoyu Cui and Huamao Jiang
   [PDF]

4. MedConvMamba: Enhancing Medical Image Classification by Integrating Convolutional Neural Networks with Mamba for Local Feature Extraction and Global Context Awareness [PDF]

MLLM&Agents

1. CorrectFlow: On-the-Spot Correction for Multimodal Reasoning with Multi-Agent Collaboration
   Xiao Dong, Pan Zhou, Xingyu Li, Zheng Chong, Yuhao Cheng, Jianxing Yu, Jian Yin, Xiaodan Liang
   [PDF]

2. ACTIONFILLER: FILL-IN-THE-BLANK PROMPTING FOR OS AGENTS
   Xiao Dong, Zijun Zhang, Xingyu Li, Yuhao Cheng, Jianxing Yu, Jian Yin, Pan Zhou, Xiaodan Liang
   [PDF]

Spatial Reasoning

1. SpatialDreamer: Incentivizing Spatial Reasoning via Active Mental Imagery
   Meng Cao^*, Xingyu Li^* (co-first author), Xue Liu, Ian Reid, Xiaodan Liang
   [PDF]

Projects

AI Pathology Image Analysis System

The AI Pathology Image Analysis System is an intelligent pathology imaging tool integrated with eye-tracking technology. It efficiently processes and analyzes pathology slide images.

Competitions&Awards

• National Undergraduate Biomedical Engineering Innovation Design Competition – National Second Prize

• ISICDM 2023 Chest X-ray Posture Assessment Competition: The Second Prize

• Northeastern University Technology Innovation Individual

Invention Patent

• A Method for Analyzing and Evaluating Medical Imaging Interpretation Skills