I was a MSc student in Stanford Vision and Learning Lab, working with Prof. Jiajun Wu and Prof. Fei-Fei Li.
I previously received my dual B.S. in Mechanical Engineering from Shanghai Jiao Tong University and Purdue University, where I was fortunate to be advised by Karthik Ramani on Human-Computer Interaction.
I support diversity, equity, and inclusion. If you would like to have a chat with me regrading research, career plans or anything, feel free to reach out! I would be happy to support people from underrepresented groups in the STEM research community, and hope my expertise can help you.
I have been working on design, fabricate, and understand tactile sensors and the rich information broght by them.
I'm also broadly interested in artificial intelligence and robotics, including but not limited to perception, planning, control, hardware design, and human-centered AI.
The goal of my research is to build agents that can achieve human-level of learning and adapt to novel and challenging scenarios by leveraging multisensory information including vision, audio, touch, etc.
We present
a method for using highly flexible, curved, passive whiskers mounted along a robot arm to gather sensory data as
they brush past objects during normal robot motion.
We introduce the OBJECTFOLDER BENCHMARK, a
benchmark suite of 10 tasks for multisensory object-centric
learning, and the OBJECTFOLDER REAL dataset, in-
cluding the multisensory measurements for 100 real-world
household objects.
We introduce SONICVERSE, a multisensory
simulation platform with integrated audio-visual simulation
for training household agents that can both see and hear.
We demonstrate SONICVERSE’s realism via sim-to-real
transfer.
We build a robot system that can see with a camera,
hear with a contact microphone, and feel with a vision-based tactile sensor,
with all three sensory modalities fused with a self-attention model.
We present
a method for using highly flexible, curved, passive whiskers mounted along a robot arm to gather sensory data as
they brush past objects during normal robot motion.
We explores a novel task ""Dexterous Grasp as You Say"" (DexGYS), enabling robots to perform dexterous grasping based on human commands expressed in natural language.
We introduce the OBJECTFOLDER BENCHMARK, a
benchmark suite of 10 tasks for multisensory object-centric
learning, and the OBJECTFOLDER REAL dataset, in-
cluding the multisensory measurements for 100 real-world
household objects.
We introduce SONICVERSE, a multisensory
simulation platform with integrated audio-visual simulation
for training household agents that can both see and hear.
We demonstrate SONICVERSE’s realism via sim-to-real
transfer.
We build a robot system that can see with a camera,
hear with a contact microphone, and feel with a vision-based tactile sensor,
with all three sensory modalities fused with a self-attention model.
Using our VRFromX system,
users can select region(s) of interest (ROI) in scanned point cloud or
sketch in mid-air using a brush tool to retrieve virtual models and
then attach behavioral properties to them.
Academic Services
Reviewer for CoRL, RAL, CHI
Teaching
Course Assistant in AA274A: Principle of Robot Autonomy, Stanford University, 2022
Course Assistant in CS231N: Deep Learning for Computer Vision, Stanford University, 2023