Wilson Yan

Hi everyone! I'm a 4th year Computer Science PhD Student at UC Berkeley. I am currently advised by Professor Pieter Abbeel as a part of Berkeley AI Research.

Email  /  Google Scholar

profile photo
Research

I am interested in generative modeling. My current research is focused on learning models for accurate video prediction and generation.

World Model on Million-Length Video and Language With Blockwise RingAttention
Hao Liu*, Wilson Yan*, Matei Zaharia, Pieter Abbeel
In Submission
project page
tweet

A vision-language model trained on million-length sequences of images, language, and video.

Motion-Conditioned Image Animation for Video Editing
Wilson Yan, Andrew Brown, Pieter Abbeel, Rohit Girdhar, Samaneh Azadi
In Submission
project page
tweet

A motion-conditioned image animation approach to video editing that enabled a diverse range of edit types

Video Prediction Models as Rewards for Reinforcement Learning
Alejandro Escontrela*, Ademi Adeniji*, Wilson Yan*, Ajay Jain, Xue Bin Peng, Ken Goldberg, Youngwoon Lee, Danijar Hafner, Pieter Abbeel
NeurIPS 2023
project page
tweet

We show that video prediction likelihoods can provide more dense reward signal for learning reinforcement learning agents

Language Quantized AutoEncoders: Towards Unsupervised Text-Imge Alignment
Hao Liu, Wilson Yan, Pieter Abbeel
NeurIPS 2023
tweet

We learn an aligned text-image token representation without text-image data, that can then be used for downstream VQA and classification tasks using in-context learning with an LLM

Temporally Consistent Transfomers for Video Generation
Wilson Yan, Danijar Hafner, Stephen James Pieter Abbeel,
ICML 2023
project page
tweet

A vector-quantized latent dynamics video prediction model that learns compressed representations to efficiently condition on long videos of hundreds of frames during both training and generation

Patch-based Object-centric Transformers for Efficient Video Generation
Wilson Yan, Ryo Okumura, Stephen James Pieter Abbeel,
project page

An video generation architecture that leverages an object-centric transformer design for more scalable and efficient video generation

VideoGPT: Video Generation using VQ-VAE and Transformers
Wilson Yan*, Yunzhi Zhang*, Pieter Abbeel, Aravind Srinivas
project page

An efficient and scalable video generation architecture that first learns a VQ-VAE to compress video data, and then learns a GPT-style transformer to model discrete latent codes

Learning Predictive Representations for Deformable Objects Using Contrastive Estimation
Wilson Yan, Ashwin Vangipuram, Pieter Abbeel, Lerrel Pinto
CoRL 2020
project page
blog post

Using contrastive methods to learn plannable representations for manipulating deformable objects such as rope and cloth

Natural Image Manipulation for Autoregressive Models using Fisher Scores
Wilson Yan, Jonathan Ho, Pieter Abbeel,
Preprint

Performing image interpolation and semantic manipulation (e.g. hair color, smiling, gender) with a PixelCN on facial images using Fisher scores as an embedding space

Learning to Manipulate Deformable Objects without Demonstrations
Yilin Wu, Wilson Yan, Thanard Kurutach, Lerrel Pinto, Pieter Abbeel,
RSS 2020
project page
blog post

Sim-to-Real transfer to learning a cloth-spreading policy on a real robot with any human demonstrations

Teaching
cs188 CS 188: Introduction to Artificial Intelligence
Undergraduate Student Instructor, CS188 Fall 2018
Undergraduate Student Instructor, CS188 Spring 2019
Head Undergraduate Student Instructor, CS188 Fall 2019

CS 294-158: Deep Unsupervised Learning
Undergraduate Student Instructor, CS294-158 Spring 2020
Website Source