Experiences


A random walk with serendipity


The Tower

Visiting Scholar at
University of California, Berkeley

May 2018 to December 2018

Developed Scalabel and interfaced the platform with various deep models.
Ongoing research includes interactive segmentation and active learning.

Georgia Tech BioQuad

Research Assistant at
Healthcare Robotics Lab

August 2016 to May 2018

Works include material recognition with tactile sensors, and pixel-level material recognition with infrared active sensing module.
Will extend to the deployment of the active sensing module to the PR2 robot after returning to Georgia Tech in Spring 2019, in order to perform object recognition in the wild.

The Love Building

CS Major at
Georgia Institute of Technology

August 2015 to May 2019

ME 1770 (Intro to Engr. Graphics & Visualization):
Learned 3D modeling and printing; had a great time prototyping my ideas.
CS 2200 (Intro to System and Networks):
Learned about computers from a system perspective; projects include constructing pipelines in simulators, processor schedulers, and transport protocols.
CS 3630 (Intro to Robotics and Perception):
Coded with the Cozmo robot to achieve different tasks in the real world; projects include localization, particle filters, and odometry control. CS 4641 (Machine Learning):
Learned classical approaches and theories of machine learning, including supervised learning, unsupervised learning, reinforcement learning, information theory and game theory. Projects include classification, optimization, and training RL agents.
CS 7643 (Deep Learning):
Topics include theories behind CNNs, RNNs, Deep RL, generative models, and visual question answering. Projects include implementing CNN from scratch, adversarial attack, style transfer and visual question answering. Reviewed relevant papers weekly and presented in class. The broad aspects of the class offers me a better understanding of the research community and a guide about how to manage deep learning projects, review papers, and present ideas to a large crowd.
CS 7476 (Advanced Computer Vision):
Topics include generative models, adversarial attack, image translation, object detection/segmentation, 3D point cloud architectures, and non-local frameworks. Reviewed two papers per week and presented in class. Presented the final project in the class poster session. This class further strengthened my skill of developing computer vision projects, writing papers, and presenting posters.

The Goddess Collage at PKU School

High school at
北京大学附属中学

August 2012 to May 2015

Founded Archer Studio, a freelancing designer club. Had great interests in graphic design, making documentary, and front-end development.