Researcher, Google
1 paper at NeurIPS 2025
learning pixel-dense multi-task representations from apparent motion