Yaoyu Zhang ⚡️
Hey there 👋
I'm Yaoyu (Allan) Zhang
PhD Student & ML Engineer
I am a third-year Ph.D. student in the Department of Mechanical and Industrial Engineering at the University of Toronto. My research focuses on robust time series analysis and classification under distributional shift, with a particular emphasis on online learning and test-time training. I am especially interested in real-world applications such as financial time series forecasting and anomaly detection.

Outside of research, I love playing badminton 🏸 — it's my favorite way to relax and stay active.
Current Projects
Robust Test-Time Adaptation under Distributional Shift

Robust Test-Time Adaptation under Distributional Shift
Research

We address the limitations of standard test-time adaptation methods, such as Tent, in scenarios where we do not have labels during testing. Our approach mitigates compounding errors and early misprediction reinforcement in pseudo-labeling by stabilizing the adaptation process under weak initial representations and distributional shifts.
Graph Attention Network Under Mixed Distributional Shift

Graph Attention Network Under Mixed Distributional Shift
Research

We aim to address the problem of multimodal or mixed distributional shift for time series classification through some graph attention network. We also try to address the stability and plasticity tradeoff while keeping the long-term dependencies.