About Me
I am a third-year Software Engineering undergraduate at Sichuan University (SCU). Under the supervision of Prof. Tao He, I previously investigated robust medical image segmentation, with a focus on Ordinary Differential Equations (ODEs) and Test-Time Adaptation (TTA).
Currently, I am a remote research intern with Prof. Yuzhang Shang at University of Central Florida (UCF), where my interests center on Vision-Language-Action (VLA) models, particularly on enhancing their reasoning and generation capabilities.
I am actively seeking Fall 2027 Ph.D. opportunities to further explore VLA and generalization.
Email: xutianqi@stu.scu.edu.cn / tianqixu692@gmail.com
🔥 News
- Jan. 20–27, 2026: Attended AAAI 2026 in Singapore on behalf of our team to deliver a poster presentation for CNM-UNet.
- Nov. 2025: Our work CNM-UNet is accepted at AAAI 2026.
📝 Publications
CNM-UNet: Continuous Ordinary Differential Equations for Medical Image Segmentation
Tianqi Xu, Yashi Zhu, Quansong He, Yue Cao, Kaishen Wang, Zhang Yi, Tao He Proceedings of the AAAI Conference on Artificial Intelligence (AAAI), 2026
| Paper | Code | Poster |
- Proposed CNM-UNet, which replaces vanilla hierarchical decoders with a single Continuous Neural Memory ODEs Block to reduce computational overhead.
- Leveraged Neural Memory ODEs (nmODEs) for continuous-time feature extraction, achieving precise data distribution modeling.
- Designed a Dual Self-updating (DUSE) strategy based on Test-time Adaptation (TTA) principles to enhance cross-domain generalization.
🎖 Honors and Awards
- 2023-2025 Merit Student Award (Consecutive 2 Years), Sichuan University.
- 2023–2024 Second-class Comprehensive Academic Scholarship (Top 6%), Sichuan University.
- 2024–2025 Third-class Comprehensive Academic Scholarship (Top 15%), Sichuan University.
Thanks to RayeRen for the template.