I am a master's student (expected graduation: June 2027) at the College of Computer Science and Software Engineering, Shenzhen University. My advisor is Prof. Huisi Wu.
Status
Applying for Fall 2027 PhD admission; available to join a research group early; currently based in Shenzhen.
Research Interests: Experience-Driven Agent Learning and Continual Adaptation; Previous Work on Efficient Temporal Memory and Stable State Updates.
Timeline
▶View Full TimelineCollapse Timeline
Present
Current Status
Applying for Fall 2027 PhD admission; available to join a research group early.
2026
CVPR 2026 Highlight + Compute Transparency Champion Award
'OSA: Orthogonal State Update with Anatomical Prior-aware Feature Enhancement for Echocardiography Video Segmentation' accepted at CVPR 2026 as a Highlight paper, and received the CVPR Compute Transparency Champion award.
2026
National Graduate Scholarship
2026
Shenzhen University Graduate Academic Scholarship (First Prize)
2026.01
Shenzhen University Academic Launch Scholarship
2025.12
National Graduate Scholarship
2025.12
Shenzhen University Graduate Academic Scholarship (First Prize)
2025
ICCV 2025 Acceptance
'GDKVM: Gated Delta Rule-based Spatio-Temporal Key-Value Memory for Echocardiography Video Segmentation' accepted at ICCV 2025.
2024.12
Shenzhen University Graduate Academic Scholarship (Special Prize)
2024.09
M.S. Admission
Joined the College of Computer Science and Software Engineering at Shenzhen University. Began research on echocardiography video segmentation.
2024.06
Henan University - Bachelor's Degree (Graduation)
2022.12
Three Good Student with Scholarship
2020.09
Bachelor's Degree (Admission)
Henan University, Data Science and Big Data Technology.
Publications
OSA: Echocardiography Video Segmentation via Orthogonalized State Update and Anatomical Prior-aware Feature Enhancement
Studying how Qwen agents learn from successful and failed WebShop tasks under limited parameter memory. I introduced Bounded State into SD-LoRA, consolidating past and new updates through a fixed-capacity recurrence to balance learning efficiency and knowledge retention. Multi-round local behavior checks passed; matched trainer comparisons averaged about 37x speedup (trainer only). Planned submission: ICLR 2028.