I was awarded the National Scholarship for the second time.
About Me
I am Zixuan Jiang (Andrew), an undergraduate student in Artificial Intelligence at Xi'an Jiaotong University. I am a member of the Young Gifted Program (2021 cohort) and the Artificial Intelligence Experimental Class (2023 cohort).
I have been admitted to the Computer Science PhD program at Shanghai Innovation Institute and the School of Computer Science, Shanghai Jiao Tong University. I will begin my doctoral studies in September 2027 under the supervision of Xie Chen, and I am currently working with the X-LANCE Lab.
My research centers on Audio-Visual Intelligence: how audio-visual systems understand and represent the world, and how they interact and collaborate with people.
I am always glad to hear about research collaborations or internships in audio-visual intelligence. Feel free to send me an email.
Outside of research, I enjoy music and playing the piano.
Research Interests
- Multimodal understanding & representation. I study how audio and video can be combined to understand what is happening and why, and how models can learn representations that capture both.
- Human–AI interaction & collaboration. I study how audio-visual systems can interpret human intent, learn from feedback, and support natural interaction.
My earlier work focused on remote sensing image interpretation, especially detailed localized captioning and open-vocabulary perception. Full publication list →
News
I have been admitted to the Computer Science PhD program at Shanghai Innovation Institute and School of Computer Science, Shanghai Jiao Tong University, starting in September 2027.
DescribeEarth: Detailed localized captioning for remote sensing images is now published in ISPRS P&RS (Vol. 242 · pp. 860–872).
AuK is now open source: unified speech generation and editing, with AuK-Flash and AuK-VAE. Paper Code Project Page
AgenticASR is now on arXiv, with the code and AASR-Bench released. Paper Code AASR-Bench
Earlier updates
OVEarth-Bench was released on arXiv with a project page for open-vocabulary Earth observation evaluation. Paper Project
The VibeXASR project, in collaboration with Jun Guo, Tao Liu, and Chonghao Cai, won the Grand Prize at the Speech OPC Innovation & Entrepreneurship Competition. Project
Our survey paper on remote sensing open-vocabulary perception has been accepted by Journal of Image and Graphics (中国图象图形学报).
DescribeEarth was made public with open-source code, dataset, and benchmark. Paper Code Dataset Benchmark
Education
- PhD in Computer ScienceIncoming
Shanghai Innovation Institute & School of Computer Science, Shanghai Jiao Tong University
- Artificial Intelligence Experimental ClassBachelor
Outstanding Talent Program, Xi'an Jiaotong University
- Young Gifted ProgramPreparatory
Xi'an Jiaotong University
Research Experience
- Shanghai Innovation Institute · Shanghai Jiao Tong University
X-LANCE Lab
Mentor: Xie Chen · Associate Professor · PhD SupervisorSpeech recognition, generation, and editing
- Xi'an Jiaotong University
School of Computer Science
Remote sensing image interpretation and open-vocabulary perception