I am a final-year PhD student in Computer Science at University of Illinois, Urbana Champaign. I am working in the Data Mining Group under Prof. Jiawei Han.
In my research, I develop data-driven, weakly supervised, and scalable techniques for knowledge discovery over massive, complex, and noisy network (graph) data. My interests span graph data mining, network data science, and applied machine learning, with a focus on designing novel graph analysis and deep learning frameworks for the effective construction, modeling, and application of real-world network data, towards tasks like conditional stucture generation, contextualized network embedding, and graph-aided recommendations.
Before joining UIUC, I received my Bachelor of Engineering Degree in Computer Science in 2014, from the Chu Kochen Honors College, Zhejiang University. I was working in the State Key Lab of CAD&CG on machine learning with Prof. Xiaofei He.
Motivated researchers/students with strong backgrounds and interests in machine learning and network data mining are welcome to get in touch for brainstorms and discussions :)
[2019.11] I will be delivering research talks in Northwestern University, Simon Fraser University, and University of British Columbia.
[2019.10] Our collaboration with LinkedIn Economic Graph Research has led to another paper accepted by WSDM 2020 (after CIKM 2019).
[2019.09] I am invited to attend and give a talk in the Great Lakes Workshop on Data Science.
[2019.09] My work with students on conditional graph generation has been accepted by Neurips 2019.
[2019.08] My two new leading research papers have been accepted by CIKM 2019 and ICDM 2019.
[2019.07] The Han Family will get together in San Francisco. Happy birthday Prof Han!
[2019.06] Our demo about Cube Networks has been accepted by KDD 2019.
[2019.05] I am working in Pinterest Lab this summer with the Applied Science team led by Prof. Jure Leskovec.
[2019.04] The first paper written by my students has been accepted by IJCNN 2019. Congratulations to the authors!
Empowered GraphSage for web-scale Pinterest graph embedding through efficient graph down-sampling and context-aware conditional aggregation on heterogeneous pin-board networks.
Developed a two-step data-driven pipeline of feature generation and metric learning for place embedding to leverage ad-hoc place attributes and noisy training data towards efficient place deduplication.
Constructed the transportation HIN (heterogeneous information network) based on DiDi's unique travel data and developed a pattern-aware HIN embedding algorithm for passenger experience prediction.
Developed a joint learning framework of user links and attributes for friend recommendation and interest targeting. Implemented a Spark pipeline and scaled up to millions of nodes and billions of edges.
Implemented data extraction and inventory analysis pipeline using flume C++. Implemented an online simulation of ads serving and an offline optimal algorithm based on max flow to analyze the inventory.