Kaiyu Zheng

Profile
  • Our paper has been accepted to AAAI 2018!

I'm a research assistant at the UW Robotics State Estimation (RSE) Lab, working with Andrzej Pronobis and Rajesh Rao since 2016. My current research topic is on learning semantic maps with topological spatial relations. We proposed Graph-Structured Sum-Product Networks (GraphSPN), a new, deep, probabilistic framwork for structured prediction. Previously, I worked on robot navigation and data collection of annotated robotic localization and sensory information. I am applying for Ph.D. for fall 2018.

My primary interests are in the intersection of machine learning and robotics. I am interested in enabling robots to learn knowledge and interact with cluttered, probabilistic environments intelligently, potentially involving collaboration with humans and other robots. I am also higly interested in applications such as medical robotics, underwater robotics, and autonomous systems. I hope to conduct research on some of these topics for my Ph.D. studies.

Currently, I study Computer Science at the University of Washington as a Master's student (B.S./M.S.). I expect to graduate in June, 2018.

Graduated from the University of Washington in June, 2017, with a B.S. degree in Computer Science with cum laude and departmental honors for my research thesis, and a minor in mathematics.

Previously, in spring 2015, I worked on device integration of Myo at the UW Movement Control Lab, advised by Vikash Kumar and Emo Todorov. In summer 2015, I interned at CME (Chicago Mercantile Exchange), in Chicago, IL, as a software engineer.

For more, please refer to my CV (Last updated on 11/13/2017). Check out my dumping ground for my projects, writings, etc.

My chinese name is 郑开宇 (simplified), or 鄭開宇 (traditional).

Conference Papers

Learning Semantic Maps with Topological Spatial Relations Using Graph-Structured Sum-Product Networks
Kaiyu Zheng, Andrzej Pronobis, Rajesh P. N. Rao
arXiv:1709.08274, 2017 [pdf] [show abstract]
Accepted at AAAI 2018. Appeared during the IROS 2017 2nd Workshop on Machine Learning Methods for High-Level Cognitive Capabilities in Robotics. [presentation]

Thesis

Learning Large-Scale Topological Maps Using Sum-Product Networks
Kaiyu Zheng
Senior Thesis, arXiv:1706.03416, 2017 [pdf] [show abstract]

Presentations

Probabilistic Semantic Mapping Using Graph-Structured Sum-Product Networks
Kaiyu Zheng, Andrzej Pronobis, Rajesh P. N. Rao
UW Computer Science & Engineering Industry Affiliates Research Day
— on 11/15/2017 in Seattl, WA, US.
[poster] [slides] [show abstract]
Learning Semantic Maps with Topological Spatial Relations Using Graph-Structured Sum-Product Networks
Kaiyu Zheng, Andrzej Pronobis, Rajesh P. N. Rao
IROS 2017 2nd Workshop on Machine Learning Methods for High-Level Cognitive Capabilities in Robotics
— on 9/28/2017 in Vancouver, BC, Canada.
[poster] [slides] [hide video]
Training Deep Probabilistic Models for Semantic Mapping with Mobile Robots
Kousuke Ariga*, Kaiyu Zheng*, Andrzej Pronobis
20th Undergraduate Research Symposium at the University of Washington, 2017
— on 5/19/2017 in Seattle, WA, US.
[poster] [show abstract]

Resources

ROS Navigation Tuning Guide
Kaiyu Zheng
arXiv:1706.09068, 2017 [pdf] [ROS site] [hide video]

More about Me

Email: kaiyuzh [at] uw [dot] edu

Github: zkytony

Dumping ground for projects, writings, etc. My interests are diverse.