Xianjie Dai

I am a Ph.D. student at the Cho Chun Shik Graduate School of Mobility, KAIST, supervised by Prof. Heye Huang.

I received my M.S. degree in Computer Science from EPFL. At EPFL, I did my semester project at the eCEO Lab with Prof. Devis Tuia and Dr. Li Mi. Before that, I obtained my bachelor's degree from PolyU.

Previously, I was a Research Assistant at SJTU, working with Prof. Peilin Zhao. I also worked as a Research Assistant at KAUST's Vision-CAIR, where I worked with Prof. Mohamed Elhoseiny and collaborated closely with Dr. Jian Ding on generative models and embodied AI, focusing on end-to-end vision-language and vision-language-action models.

Email / Scholar / LinkedIn / GitHub

Research

My current work focuses on in-context learning and generalization in embodied AI. More broadly, I am interested in multimodal learning, vision-language models, representation learning, and reliable decision-making systems.

Publications

	ContextFlow: In-Context Flow Matching for Robot Manipulation Jian Ding, Xianjie Dai, Roei Herzig, Nussair Hroub, Jinjie Mai, Dengxin Dai, Bernard Ghanem, Mohamed Elhoseiny European Conference on Computer Vision (ECCV), 2026 paper / project page * Equal contribution.
	Knowledge-aware Text-Image Retrieval for Remote Sensing Images Li Mi, Xianjie Dai, Javiera Castillo-Navarro, Devis Tuia IEEE Transactions on Geoscience and Remote Sensing (TGRS), 2024 paper
	ConVQG: Contrastive Visual Question Generation with Multimodal Guidance Li Mi, Syrielle Montariol, Javiera Castillo-Navarro, Xianjie Dai, Antoine Bosselut, Devis Tuia AAAI Conference on Artificial Intelligence (AAAI), 2024 paper / project page

Education

Korea Advanced Institute of Science and Technology (KAIST)
Ph.D. student, Cho Chun Shik Graduate School of Mobility
Advisor: Prof. Heye Huang

École Polytechnique Fédérale de Lausanne (EPFL)
M.Sc. in Computer Science

The Hong Kong Polytechnic University
B.Eng. (Hons) in Electronic and Information Engineering

Misc

My research experience spans embodied AI, vision-language learning, biomedical imaging, and multimodal representation learning.

Website template adapted from Jon Barron's academic website.