Graph Learning
TSD-GCN
A topology-learnable static–dynamic graph convolutional network that captures complementary functional-connectivity patterns for brain disorder detection.
Research
Brain Imaging & Representation Learning
I am a third-year graduate student working on intelligent diagnosis of brain disorders with resting-state functional MRI. My research explores brain network modeling and representation learning, with a focus on cross-site generalization, static–dynamic graph learning, and causal representations of brain dynamics.
Research Highlights
Causal Learning
Ca-MIL
A multi-instance causal representation framework that models temporal variations in brain dynamics and improves disease classification from rs-fMRI.
Cross-Site Learning
GACRL
A geometry-aware causal representation learning framework that preserves brain-network geometry on Riemannian manifolds and learns both domain-invariant and domain-dependent causal representations.
Domain Adaptation
DDA-LRR
A discriminative domain adaptation method that uses low-rank representation, orthogonal reconstruction, and source-domain supervision for multi-site autism spectrum disorder identification.
Publications
Topology-Learnable Static-Dynamic Graph Convolutional Network for Brain Disorder Detection with Functional MRI
Second author · IEEE Transactions on Neural Systems and Rehabilitation Engineering
Discriminative Domain Adaptation via Low-Rank Representation for Multi-site Autism Spectrum Disorder Identification
First author · Journal of Data Acquisition and Processing, 38(4), 886–897
Manuscripts
Multi-Instance Causal Representation Learning of Brain Dynamics for Disease Detection
Second author · Minor revision at Medical Image Analysis
Education
M.Sc. in Computer Science
NUIST · 2024–2027 (expected)
GPA: 4.15/5 · Top 4%
B.Eng.
NUIST · 2020–2024
Research Interests
- Medical Image Analysis
- Brain Network and Time-Series Analysis
- Domain Adaptation and Causal Representation Learning