Jiaxuan Li

李嘉烜

I am a Research Scientist at SB Intuitions Corp., where I work on large language models and vision-language models.

I received my Ph.D. in Information Science and Technology from the University of Tokyo in 2025 under the supervision of Prof. Hideki Nakayama at the Nakayama Laboratory. Previously, I earned my M.Eng. from Shanghai Jiao Tong University under the supervision of Prof. Yuye Ling, and my B.Eng. from Shanghai University under the supervision of Prof. Zhi Liu.

My research interests include multimodal reasoning, hallucination mitigation, knowledge grounding, and trustworthy AI.

I am open to research collaborations on multimodal and trustworthy AI. Please feel free to get in touch.

News

Aug 2026Our paper “Do VLMs Share Safety Neurons Across Modalities?” was accepted to EMNLP 2026 (Main).
Apr 2025Joined SB Intuitions as a Research Scientist, focusing on responsible AI.
Mar 2025Received my Ph.D. in Information Science and Technology from the University of Tokyo.
Oct 2024Selected as a DAAD AInet Fellow by the German Academic Exchange Service (DAAD).
Aug 2024Gave an invited talk on EVCap at MIRU 2024 in Kumamoto.
May 2024Received a Google Student Travel Grant for East Asia.
Apr 2024Selected to attend the International Computer Vision Summer School (ICVSS 2024) in Sicily, Italy.
Feb 2024One paper was accepted to CVPR 2024.
Jul 2023One paper was accepted to ICCV 2023.
Apr 2023Selected as a Japan Society for the Promotion of Science (JSPS) Research Fellow (DC2).
Apr 2022Selected for the SPRING GX Fellowship, supported by the Japan Science and Technology Agency (JST).

^* Equal contribution

	Do VLMs Share Safety Neurons Across Modalities? Jiaxuan Li^, Jiahao Zhang^, MinhDuc Vo, Huy H. Nguyen, Pride Kavumba, Koki Wataoka EMNLP 2026 (Main)
	NEMO: Can Multimodal LLMs Identify Attribute-Modified Objects? Jiaxuan Li, Junwen Mo, MinhDuc Vo, Akihiro Sugimoto, Hideki Nakayama arXiv:2411.17794 paper website code
	EVCap: Retrieval-Augmented Image Captioning with External Visual-Name Memory for Open-World Comprehension Jiaxuan Li^, Duc Minh Vo^, Akihiro Sugimoto, Hideki Nakayama CVPR 2024 paper website code
	Partition-And-Debias: Agnostic Biases Mitigation via a Mixture of Biases-Specific Experts Jiaxuan Li, Duc Minh Vo, Hideki Nakayama ICCV 2023 paper code

Japan Society for the Promotion of Science (JSPS), Grant-in-Aid for JSPS Fellows, JPY 2M, Apr 2023–Mar 2025
National Institute of Informatics (NII), NII Open Collaborative Research (Researcher), JPY 1.2M, Jul 2024–Mar 2025
National Institute of Informatics (NII), NII Open Collaborative Research (Researcher), JPY 1M, Jul 2023–Mar 2024
Japan Science and Technology Agency (JST), SPRING GX Fellowship, JPY 620K, Apr 2022–Mar 2023

AInet Fellow, German Academic Exchange Service (DAAD), 2024
Final-Year Special Allowance for JSPS Research Fellows, Japan Society for the Promotion of Science, 2024
Google Student Travel Grant for East Asia, Google, 2024
Special Committee Award, the 30th Annual Meeting of the Association for Natural Language Processing, 2024
JSPS DC2 Research Fellowship for Young Scientists, Japan Society for the Promotion of Science, 2023
SPRING GX Fellowship, Japan Science and Technology Agency, 2022
Outstanding Graduate, Shanghai Jiao Tong University, 2022
COSCO Shipping Scholarship, Shanghai Jiao Tong University, 2021
Second Prize, “Huawei Cup” 17th China Post-Graduate Mathematical Contest in Modeling, 2020
Academic Scholarship, Shanghai University, 2016

Invited Talk, Knowledge and Information Research Team, AIST, Japan, Feb 2025
Invited Talk, MIRU, Kumamoto, Japan, Aug 2024
Poster Presentation, International Research Center for Neurointelligence (IRCN), The University of Tokyo, Japan, Jul 2024
Invited Talk, Knowledge and Information Research Team, AIST, Japan, Jul 2024
International Computer Vision Summer School (ICVSS), Sicily, Italy, Jul 2024
Reviewer at IEEE TIP, NeurIPS, CVPR, ECCV, ACL ARR, AAAI, BMVC, ACCV

Teaching Assistant, SJTU AI2614: Digital Signal and Image Processing, Spring 2021
Teaching Assistant, SJTU EE367: Fundamentals of Communication Circuits, Spring 2020