Electrical and Computer Engineering
Office: Gross Hall, Durham, NC 27710
I am currently a postdoctoral associate at Duke University, where I work with Prof. Lawrence Carin.
Before that, I worked with Prof. Rebecca J. Passonneau at both Penn State University and Columbia University, in a list of projects with a focus on applying Artificial Intelligence for Education.
I got my Ph.D. from Tsinghua University in July, 2017. My advisor is Prof. Gerard de Melo. I was a member of the Web Mining and Language Technology group at Tsinghua University, part of the Institute for Interdisciplinary Information Sciences (IIIS), which is led by Turing award winner Andrew Yao. Before that, I received my Bachelor of Computer Science from Beijing University of Posts and Telecommunications, with two publications (indexed by SCI, Impact factor ~3) in the area of quantum information and top ~1% of GPA rank.
My research focuses on a wide range of topics in Natural Language Processing (especially Text Summarization, Open Information Extraction, Knowledge Graph, Machine Translation), Deep Learning, and the intersection of AI and HCI. For example,
Text Summarization: We propose PEAK, the very first method for automatically evaluating the quality of a summary using the Pyramid method. PEAK shows great promise for automated scoring and assessment of manual or automated summaries, opening up the possibility of widespread use in the education domain and in information management.
Machine Translation: To overcome the data scarcity problem for resource-scarce language pairs, we introduce a joint training algorithm for pivot-based neural machine translation, which leads to significant improvements over independent training across various languages.
Deep Learning: We decouple the backpropagation by breaking the backward lock using delayed gradients, which accelerates the training of deep learning models significantly without loss of accuracy.
Human-Computer Interaction: The staggering amounts of content readily available to us via digital channels can often appear overwhelming. So we create HiText, a simple yet effective way of dynamically marking parts of a document in accordance with their salience. HiText relies on a deep learning-based sentence ranking method that results in marked increases in user satisfaction and reading efficiency.
We provides a Java implementation of PEAK, the first tool to automatically generate the pyramid models that enables the whole automatical assessment of both human and machine summaries.
[Enter PEAK's homepage]
HiText: Text Reading with Dynamic Salience Marking.
Neural Machine Translation with Pivot Languages.
Wise Crowd Content Assessment and Educational Rubrics.
International Journal of Artificial Intelligence in Education, 10.1007/s40593-016-0128-6, 2016.
Research on the Key Management for AMI System Based on IEC62056 Standard.
Electrical Measurement and Instrumentation, Issue No. 2, 2016.
Semi-loss-tolerant Strong Quantum Coin-flipping Protocol Using Quantum Non-demolition Measurement.
Quantum Information Processing, 10.1007/s11128-014-0747-5, 2014.
Semi-loss-tolerant Strong Coin Flipping Protocol Using EPR Pairs.
Quantum Information and Computation, Vol. 12 No. 5 and 6, 2011.
Three-party secret key management method based on IEC62056 data transmission safety standard.
Filing date: Nov 6, 2014
2017.4 – 2017.5, visiting student, the Department of Computer Science, Rutgers University
Supervisor: Prof. Gerard de Melo
2016.5 – 2016.9, visiting student, the Human Language Technology and Pattern Recognition Group, Computer Science Department, RWTH Aachen University
Supervisor: Prof. Dr.-Ing. Hermann Ney
CSE 597 Spring 2018, Natural Language Processing, Penn State University
Co-teach with Prof. Rebecca J. Passonneau
Spring 2014 – 2015, Speech Science, Technology and Innovative Applications, Tsinghua University
TA of Instructors: Prof. P.C. Ching, Prof. Tan Lee, Prof. Ken Ma, Prof. Helen Meng, and Prof. William S.-Y. Wang, all from the Chinese University of Hong Kong, and Prof. Gerard de Melo, from Tsinghua University
Fall 2013 – 2014, Hot Topics in Computational Biology, Tsinghua University
TA of Instructor: Prof. Jianyang (Michael) Zeng
Program Committee Member:
NAACL(2018), IJCAI-ECAI(2018), ACL(2018)
Assigned to LINYANG Electronics Co., Ltd. by Tsinghua University, Jul. 2014
Topic: Data Encryption for Advanced Metering Infrastructure System
One publication in this area, which is not related to my PhD topic.
Awarded Outstanding Individual in Graduates' Summer Internship of Tsinghua University, 2014
Host for IIIS New Year Party 2013, 2014
You would never know the host is now very hungry.
Assistant for Graduate Students of IIIS, 2013
When organizing this autumn outing, some students wanted to go hiking but some only wanted to go to the KTV or play games, and the other preferred to eat delicious food just as me. Finally I found a place where we made all the ideas come true.
We organized a successful Doctoral Consortium with PhDs from both Computer Science and Physics.
This is a chorus competition host by Tsinghua University, but the judgers come from the top singers in China. As the assistant, I'm proud of the prize we won!
I love basketball very much and I can throw at least 4 balls given 5 ones into the basket. BTW, could you find me in the second pictures? (Hint: just find a girl out of all the other men players in the football team.) It's the first time the team won a competition!
Vice Minister of the Academic Ministry of Tsinghua University Graduate Association, 2013
League Secretary of IIIS Grad2, 2012
I'm one of the editors.Awarded Outstanding Student Leader of Tsinghua University, 2014
Interests: Go (from the age of five), Piano, Guitar, Yoga, and my photography is at 500px.
Credits: Teaching icon from Icons8.