I am a PhD student at Mohamed bin Zayed University of Artificial Intelligence (MBZUAI), advised by Prof. Ivan Laptev and Dr. Fabio Pizzati.
Previously, I earned an MSc in Computer Science from the Technical University of Munich and a BEng in Computer Science from Dalian University of Technology.
My current work focuses on physically plausible generative modeling and robot learning.
In my spare time, I enjoy cycling and traveling.
(* and † indicate co-first authorship and corresponding author, respectively.)
MessyKitchens provides high-quality synthetic training data and real-world test scenes for contact-rich kitchen reconstruction, with improved registration and reduced inter-object penetration. Built on it, MOD achieves state-of-the-art reconstruction and physical metrics.
This work introduces TARGO, a benchmark dataset for target-driven grasping under occlusions in cluttered environments, along with TARGO-Net to improve grasping performance in these physically challenging settings.
Given partial deformation, we applied a decoder-only transformer to autoregressively generate diverse and complete deformation.
We refined motion strategies for a curling robot and implemented a neural network regression model to predict the relationship between robot velocity and current position.
We developed a paper counting device using least squares fitting to correlate paper quantity and capacitance measurements from the FDC2214 sensor.
We developed a target tracking system that controls a vehicle to follow a predefined path and parses data exchanged with a millimeter-wave radar.
