TPAMI
ABIGX: A Unified Framework for Explainable Fault Detection and Classification
IEEE Transactions on Pattern Analysis and Machine Intelligence (TPAMI), 2026.
About Me
I am a Schmidt AI in Science Postdoctoral Fellow at Cornell University. My research focuses on trustworthy and explainable AI: understanding why models make decisions, when and how they fail, and how to make their predictions more interpretable, robust, and reliable.
My methodological work covers feature attribution, counterfactual explanations, adversarial robustness, formal verification, and uncertainty-aware evaluation. My explainable AI research has been published in IEEE Transactions on Pattern Analysis and Machine Intelligence (TPAMI).
I also apply computer vision, foundation models, and geospatial AI to industrial fault diagnosis, urban energy systems, and scientific discovery. I received my Ph.D. from Zhejiang University in 2023 and was previously a Research Fellow at Nanyang Technological University.
Research Directions
- Trustworthy & Explainable AI: Feature attribution, counterfactual explanations, adversarial robustness, and formal verification.
- Industrial AI & Smart Manufacturing: Anomaly detection, process monitoring, and explainable fault diagnosis.
- AI for Energy & Urban Systems: Geospatial AI, rooftop photovoltaics, remote sensing, and urban digital infrastructure.
- Computer Vision & AI for Science: Foundation models, monocular depth estimation, and 3D plant reconstruction.
Recent Highlights
TPAMI
Applied Energy
AI-enabled remote sensing reveals two urban rooftop PV adoption pathways: household vs. large-roof
Applied Energy, 426, 128766, 2026.
Comp. & Elec. Ag.
MIRAGE: Biomechanically interpretable 3D generation and reconstruction of maize plants
Computers and Electronics in Agriculture, 248, 111774, 2026.
TPAMI
IG²: Integrated Gradient on Iterative Gradient Path for Feature Attribution
IEEE Transactions on Pattern Analysis and Machine Intelligence (TPAMI), 2024.
More publications · Research posts · Curriculum Vitae
Based on Minimal Light. Powered by Jekyll · GitHub Pages.