Assistant Engineer

Ye Liu

Contact

Information

Department of Computer Science, 851 S. Morgan (M/C 152) Email: *******@***.*** University of Illinois at Chicago, Chicago, IL 60607 Mobile: 312-***-**** Education Ph.D. Candidate, Computer Science 2016 - 2021(expected) University of Illinois at Chicago Chicago, IL

Advisor: Prof. Philip S. Yu

M.S., Electronic and Computer Engineering 2014 - 2016 University of Illinois at Chicago Chicago, IL

B.S., Computer Science and Technology 2011 - 2015

Northeastern University Shenyang, China

Project Question Answering/Commonsense Reasoning

Incorporated external knowledge (textual or knowledge graph information) to reason beyond the given context to improve the performance of MRC/commonsense reasoning.

Generated the well-formed question from the ill-formed question to bene t QA model. Related paper: [1], [2], [6], [7]

Natural Language Generation

Used reinforcement learning/ generative adversarial network/ variational auto-encoder to improve the generation performance in several generation tasks.

Augmented the pre-trained seq2seq model (BART) with the knowledge graph to better capture the relation between entities and enrich the generated sentence.

Extended pre-trained language model (BERT/RoBERTa) to the multiseq2seq model to deal with the zero-shot multi-source generation with knowledge transferring from single-source generation.

Proposed structural-aware non-autoregressive transformer to speed up the decoding process mean- while considering the dependence structure between tokens. Related paper: [1], [2], [3], [8], [9], [10]

Graph/Knowledge Graph Embedding

Considered the multi-view multi-graph information to improve the graph representation. Related paper: [4], [7]

Publication 1. Ye Liu, Chenwei Zhang, Xiaohui Yan and Philip S. Yu. \Generative Question Re nement with Deep Reinforcement Learning in Retrieval-based QA System." In Proceedings of CIKM, 2019. 2. Ye Liu, Tao Yang, Zeyu You, Wei Fan and Philip S. Yu. \Commonsense Evidence Generation and Injection in Reading Comprehension." In Proceedings of SIGDIAL, 2020. 3. Jianguo Zhang, Pengcheng Zou, Zhao Li, Yao Wan, Ye Liu, Xiuming Pan, Yu Gong and Philip S. Yu. \Product Title Re nement via Multi-Modal Generative Adversarial Learning." In Proceedings of NIPS Workshop, 2018.

4. Ye Liu, Lifang He, Bokai Cao, Philip S. Yu, Ann B Ragin and Alex Leow. \Multi-View Multi- Graph Embedding for Brain Network Clustering Analysis." In Proceedings of AAAI, 2018. 5. Ye Liu, Jiawei Zhang, Chenwei Zhang and Philip S. Yu. \Data-driven Blockbuster Planning on Online Movie Knowledge Library." In Proceedings of IEEE BigData, 2018. 6. Ye Liu, Shaika Chowdhury, Chenwei Zhang, Cornelia Caragea and Philip S Yu. \Interpretable Multi-Step Reasoning with Knowledge Extraction on Complex Healthcare Question Answering." https://arxiv.org/abs/2008.02434

7. Ye Liu, Yao Wan, Lifang He, Peng Hao and Philip S Yu. \KG-BART: Knowledge Graph- Augmented BART for Generative Commonsense Reasoning." http://arxiv.org/abs/2009.12677 Un- derview of AAAI, 2021.

Manuscript 8. Jingfeng Zhang, Haiwen Hong, Zhi Li, Yin Zhang, Yao Wan, Ye Liu and Yulei Sui. \Multi- Lingual Code Semantics Disentangle via Variational Auto-Encoder with Cross-Training." Underview of AAAI, 2021.

1

9. Ye Liu, Yao Wan, Jianguo Zhang, Wenting Zhao and Philip S Yu. \Enriching Non-Autoregressive Transformer with Syntactic and Semantic Structures for Neural Machine Translation." Underview of EACL, 2020.

10. Ye Liu, Zhiyong Teng, Jibin Gong, Yao Wan, Jiawei Zhang, Jie Tang and Philip S Yu. \Zero- Shot Course Concept Summarization from Multi-Video Captions on MOOC Platform." Underview of WWW, 2021.

Experience Research Intern

Tencent LLC American, Palo Alto, USA June 2019 - August 2019

Research work in commonsense reasoning by importing explainable background knowledge.

Investigated the multi-document multi-hop question answering on reading comprehension. Research Intern

I ytek Co.,Ltd, Big Data Research Center, China May 2018 - July 2018

Applied Deep Q-learning method to solve the Real-Time Bidding problem in the advertisement.

Used the deep learning method to solve the CTR forecast on the unbalanced data. Engineer Intern

Shanghai Googol Ouchen Intelligent Technology Co., Ltd., China July 2015 - August 2015

Built radical Chinese characters Library and programmed manipulator to write Chinese letter. Undergraduate Research Assistant

Northeastern University, China Dec 2012 - May 2014

Developed Andriod App for online shopping and test the software products in Northwestern University, Shenyang University and other universities. Honors and

Awards AAAI-18 Scholarship 2018

China Scholarship Council of Outstanding Undergraduate Transfer Students 2015

Bronze Prize, National Students Creative Youth Entrepreneurship 2014

Second Prize, Challenge Cup Collegiate Business Competition of Liaoning Provience 2014

Five times of Second Prize or Third Prize, Scholarship of NEU 2011-2014

Three times of Outstanding Students of NEU 2011-2014 Invited Talk Amazon Graduate Research Symposium 2019 Generative Question Re nement with Deep Reinforce Learning. Reviewer IEEE Transactions on Cybernetics

Courses Deep Learning for Natural Language Processing, Advanced Topics in Machine Learning, Introduction to Machine Learning, Advanced Data Mining, Data Mining and Text Mining, Computer Algorithms I, Database Systems

Skill Languages: English, Chinese

Machine learning tools and libraries: Numpy, Scikit-Learn, Tensor ow, Keras Programming Languages: Python, C/C++, Matlab, Java, SQL Operating Systems: Unix/Linux, Windows

Teaching

Experience

Graduate Teaching Assistant

Department of Computer Science, University of Illinois at Chicago

Introduction to Machine Learning

Programming Practicum

Mathematical Foundations of Computing

C/C ++ Programming for Engineers with MatLab

Introduction to Computing and Programming

2

Contact this candidate