I work on teaching machines a high-level creative understanding of the dynamic real world, which I believe requires learning from video without manual supervision. I am also interested in language, machine learning, image synthesis, and interaction as they relate to this.

I am a first-year PhD student at Berkeley AI Research, advised by Alexei Efros. I also work with Chen Sun, Jiajun Wu, and Cordelia Schmid at Google Research.

I graduated from Columbia University with a BS in computer science, where I was lucky to be introduced to computer vision and advised by Carl Vondrick.

RESEARCH

  • Video Representations of Goals Emerge from Watching Failure
    Dave Epstein, Carl Vondrick
    PaperProject Page
  • Oops! Predicting Unintentional Action in Video
    Dave Epstein, Boyuan Chen, Carl Vondrick
    PaperProject PageTalk
    CVPR 2020
  • Learning to Learn Words from Visual Scenes
    Dídac Surís*, Dave Epstein*, Heng Ji, Shih-Fu Chang, Carl Vondrick
    PaperProject PageTalk
    ECCV 2020
  • What's Missing from Self-Supervised Representation Learning?
    Dave Epstein*, Yiliang Shi*, Eugene Wu, Carl Vondrick
    Paper
  • NEUZZ: Efficient Fuzzing with Neural Program Learning
    Dongdong She, Kexin Pei, Dave Epstein, Junfeng Yang, Baishakhi Ray, Suman Jana
    Paper
    IEEE S&P 2019

TEACHING

At Columbia
Advanced Computer Vision
Teaching Assistant
COMS 6998 (Spring 2019)
Data Structures and Algorithms
Head Teaching Assistant
COMS 3134 (Fall 2017- Summer 2019)