01
Research Scientist ·
Also a painter — commissions open →
evonneng.28 [at] gmail [dot] com
01 About
I build multimodal and generative models for videos, 3D animation, and robotics. My research focuses on perceiving and generating human behavior from text, audio, and visual signals.
I spent 5 years as a Ph.D. student in UC Berkeley's EECS program, co-advised by Prof. Trevor Darrell and Prof. Angjoo Kanazawa. During my PhD I focused on social and conversational AI: modeling how people move, listen, and interact so machines can engage in meaningful social interactions. Before all that jazz, I was advised by Prof. Kristen Grauman at UT Austin working on ego-exo human reconstruction.
News
2026Giving a talk at ICMI — find me at GENEA Workshop.
2024Graduated from UC Berkeley — Ph.D. in EECS (AI/CV).
2024Another year, another round of ASI@ECCV 2024.
2023The 2nd annual Artificial Social Intelligence Workshop accepted at ICCV 2023.
2022Co-organized the first Artificial Social Intelligence Workshop at CVPR 2022.
2021My piece 'Penny' selected for display on the 3rd floor of Sproul Hall, UC Berkeley.
2020A fun feature in Women in Computer Vision, CVPR Daily.
02 Selected Work
02
03
Embody 3D: a large-scale multimodal motion and behavior dataset
arXiv 202504
From Audio to Photoreal Embodiment: Synthesizing Humans in Conversations
CVPR 202405
06
07
Learning to Listen: Modeling Non-Deterministic Dyadic Facial Motion
CVPR 202208
Body2Hands: Learning to Infer 3D Hands from Conversational Gesture Body Dynamics
CVPR 202109
You2Me: Inferring Body Pose in Egocentric Video via First and Second Person Interactions
CVPR 2020 · Oral03 Art Commissions open →
A running series of the places I've been — my own "style transfer" of photos into paint, and a way to commemorate the spots I've been lucky to visit. People ask if they're AI-generated, which I suppose is a compliment these days. I also take on commissions: if you'd like a custom piece, get in touch.
