YuXuan (Andrew) Liu
I am a PhD student in CS at UC Berkeley in Pieter Abbeel's lab. Previously, I was a Masters and undergraduate student at Berkeley also in Pieter's lab in BAIR. Currently I am at covariant.ai.

Research
Self-Supervised Instance Segmentation by Grasping
YuXuan Liu, Xi Chen, and Pieter Abbeel
[pdf]
[arxiv]
Distributional Instance Segmentation: Modeling Uncertainty and High Confidence Predictions with Latent-MaskRCNN
YuXuan Liu, Nikhil Mishra, Pieter Abbeel, and Xi Chen
In the International Conference on Robotics and Automation (ICRA), 2023
[pdf]
[arxiv]
[dataset]
[code]
Autoregressive Uncertainty Modeling for 3D Bounding Box Prediction
YuXuan Liu, Nikhil Mishra, Maximilian Sieb, Yide Shentu, Pieter Abbeel, and Xi Chen
In the European Conference on Computer Vision (ECCV), 2022
[pdf]
[arxiv]
[dataset]
[code]
[blog]
Self-Consistent Trajectory Autoencoder: Hierarchical Reinforcement Learning with Trajectory Embeddings
John D. Co-Reyes*, YuXuan Liu*, Abhishek Gupta*, Benjamin Eysenbach, Pieter Abbeel, Sergey Levine
In the International Conference on Machine Learning (ICML), 2018
[pdf]
[arxiv]
[video]
[code]
Meta-Reinforcement Learning of Structured Exploration Strategies
Abhishek Gupta, Russell Mendonca, YuXuan Liu, Pieter Abbeel, Sergey Levine
In the Advances in Neural Information Processing Systems (NeurIPS), 2018
[pdf]
[arxiv]
[code]
Imitation from Observation: Learning to Imitate Behaviors from Raw Video via Context Translation
YuXuan Liu*, Abhishek Gupta*, Pieter Abbeel, Sergey Levine
In the International Conference on Robotics and Automation (ICRA), 2018
[pdf]
[arxiv]
[video]
[code]
Learning Invariant Feature Spaces to Transfer Skills with Reinforcement Learning
Abhishek Gupta*, Coline Devin*, YuXuan Liu, Pieter Abbeel, Sergey Levine
In the International Conference on Learning Representations (ICLR), 2017
[pdf]
[arxiv]
[video]
Projects
Robo Call Screener - Phone AI
Filter robocalls from humans with live transcription and responses at your fingertips- Have our robot answer your robocalls so you don't have to
- Screen humans from robots without having to pick up the phone
- Live transcription and text responses give you the flexibility to respond when you can't talk
Temporal Convolutional Policy Networks [pdf]
We show that temporal convolutional policies can be effective for partially observed tasks
including classical control and pong.
ARCode
[project page]
[video]
We designed an audio transmission protocol to send messages, links, and pictures using the Fast-Fourier Transform.
iSwipe [github]
First gesture-based keyboard for iOS using a dynamic programming algorithm to match words with vector angle differences
MindChat
[project page]
[video]
We learn EEG states to encode brain activity and transmit messages using Huffman encoding.
Dreamboard [github]
Powerful iOS theming platform with interactive widgets, custom script parser, and over 11 million downloads in Cydia.
Intellifit
[project page]
We built an iOS fitness tracker that collects real-time motion data and classifies activities using dynamic time warping.
H-Auth
[project page]
We learn handwriting styles with a neural network and built an authentication platform based on handwriting.
Instatag
[project page]
[video]
We automatically generate Instagram tags using image classification and real-time Instagram data.
Meme Generator
[project page]
We trained an image captioning model to generate meme captions conditioned on input images.
Scribe
[project page]
[video]
We applied speech recognition on lecture videos to match lectures with corresponding textbook material using tf-idf.
HouseRank
[github]
We trained a ranking system on housing preferences to filter listings on Craigslist.
Kerve
[project page]
We implemented a pose classification algorithm with the Kinect for learning Yoga.
MineCraft Mini
[github]
Minecraft clone in Java OpenGL with physics simulation, water flood fill, and perlin noise random map generation.
HoloScan
[project page]
We built a real-time facial detection and recognition system that can identify people you are looking at.
SmartRoomba
[project page]
[video]
We hacked a Roomba to follow you with real-time face detection.
Watchpoint
[project page]
We built a smartwatch app that alerts you when your phone is out of a certain range.