Paper: World Model on Million-Length Video and Language With Blockwise RingAttention
Paper: Motion-Conditioned Image Animation for Video Editing
Paper: Video Prediction Models as Rewards for Reinforcement Learning (NeurIPS 2023)
Paper: Language Quantized AutoEncoders: Towards Unsupervised Text-Image Alignment (NeurIPS 2023)
Paper: Temporally Consistent Transformers for Video Generation (ICML 2023)
Paper: Patch-based Object-centric Transformers for Efficient Video Generation
Paper: VideoGPT: Video Generation using VQ-VAE and Transformers
Paper: Learning Predictive Representations for Deformable Objects Using Contrastive Estimation (CoRL 2020)
Paper: Natural Image Manipulation for Autoregressive Models using Fisher Scores
Paper: Learning to Manipulate Deformable Objects without Demonstrations (RSS 2020)
Education
PhD from UC Berkeley, advised by Professor Pieter Abbeel.
Background
Senior research scientist at Google DeepMind. Current research focuses on world models for control.
Miscellany
Served as an Undergraduate Student Instructor and Head Undergraduate Student Instructor for CS 188: Introduction to Artificial Intelligence, and as an Undergraduate Student Instructor for CS 294-158: Deep Unsupervised Learning.