I am currently working on my master degree of science and am also looking for research and internship projects. This fall, I just finished my undergraduate from Zhejiang University/University of Illinois at Urbana-Champaign Institute. I want to thank all my friends and professors for all the help during such a wonderful time!
My research interest lies in natural language processing, knowledge graphs, and image recognition. I was involved in two research projects about Knowledge Graph and had learned much from these valuable experiences.
In my free time, I enjoy watching movies, playing erhu, and swimming.
You can find my CV here or at the CV page of this website.
- July, 2021: Our paper Structured Self-Supervised Pretraining for Commonsense Knowledge Graph Completion was accpted by Transactions of the Association for Computational Linguistics!
- November, 2020: Invited as a reviewer in Thirty-Fifth AAAI Conference on Artificial Intelligence!
- June, 2020: Our work Knowledge Graph Construction for Intelligent Maintainance was selected as top 20 of Student Academic Achievements (Qizhen Cup) at Zhejiang University.
- September, 2019: Our paper Knowledge Graph Construction for Intelligent Maintainance of Power Plant was accepted by IEEE ICEBE 2019!
- November, 2018: My team in Social Partice 2018 Zhejiang University was awarded University-level Outstanding Team (top 10) as well as Chancheng Scholarship. Our work includes volunteer teaching in rural areas and investigating local enterprise in three contries: Sri Lanka, Indonesia and Cambodia.
Structured Self-Supervised Pretraining for Commonsense Knowledge Graph Completion (accpted by TACL)
Jiayuan Huang, Yangkai Du, Shuting Tao, Kun Xu, Pengtao Xie
Knowledge Graph Construction for Intelligent Maintainance of Power Plant (IEEE ICEBE 2019)
Yangkai Du, Jiayuan Huang, Shuting Tao, Hongwei Wang
A Deep-Learning Based Framework for Construction and Reasoning of Knowledge Graph from Power Plant Operation Report (in submission)
Tingyu Xie, Jiayuan Huang, Yangkai Du, Shuting Tao, Qi Li, Hongwei Wang
Commonsense Knowledge Graph Complementation with High-order Structures
Advisor: Pengtao Xie, UC San Diego
We propose several approaches which leverage the high-order structure in Commonsense Knowledge Graphs (CKGs) to capture the high-order relationships between concepts. Specifically, we consider four types of structures: 1) concepts have long-range relations; 2) multiple paths exist between a pair of source and target concept; 3) each concept has multiple inbound relations and outbound relations with other concepts; and 4) each concept is involved in a local graph and initiates a path.
Based on these structures, we propose four pretraining methods of the concept generation models: 1) pretraining on individual paths; 2) path-to-path pretraining; 3) router pretraining; and 4) graph-to-path pretraining. Experiments on two datasets demonstrate the effectiveness of our methods. The paper has been accpted by TACL.
Knowledge Graph Construction for Intelligent Maintainance
Advisor: Hongwei Wang, Zhejiang University
We proposed a framework to build knowledge graphs for the power plants to extract knowledge and data from a large number of non-structured power plant maintenance reports and build knowledge graphs based on the relation between knowledge entities. Our paper gets published at the conference IEEE ICEBE 2019, and we are awarded as Best Paper of Conference.
FPGA based Game Design
Advisor: Chushan Li, Zhejiang University
We designed and implemented a game based on FPGA board, including game logic and drivers for peripheral equipment in C and FPGA programming in System Verilog. Our game is an advanced version of the original Pacman, which is added multi-player mode and Ghost AI.
Linux System Design
Advisor: Steven S. Lumetta, University of Illinois Urbana-Champaign
We implemented a x86-based operating system from scratch, including memory paging and segmentation, drivers of peripheral equipment and multi-process scheduling.