Ma, Xuezhe, Yingkai Gao, Zhiting Hu, Yaoliang Yu, Yuntian Deng, and Eduard Hovy. "Dropout with Expectation-linear Regularization.", ICLR 2017.
Hu, Zhiting, Poyao Huang, Yuntian Deng, Yingkai Gao, and Eric P. Xing. "Entity Hierarchy Embedding." In Proceedings of the 53rd Annual Meeting of the Association for Computational Linguistics and the 7th International Joint Conference on Natural Language Processing (ACL-IJCNLP), vol. 1, pp. 1292-1300. 2015.
GAO Yingkai, WANG Anran, WANG Xiang, TAN Chuanqi, GUO Lixuan, YANG Xiaoyuan: Generalization and Application of Leibniz Formula and Integration by Parts. Studies in College Mathematics. (2013):006.
Projects
Scientific Table Search Using Keyword Queries [data]
We propose a probabilistic framework that is capable of incorporating structural and semantic information from both query and table sides. We consider the structure of tables and give different emphasis to table components. On the query side, thanks to external knowledge such as knowledge bases and ontologies, key concepts are extracted and used to build structured queries, and target quantity types are identified and used to expand original queries. We also construct TableArXiv, a high quality dataset with 105 queries and corresponding relevance judgements for scientific table search. Conducted on TableArXiv, experiments using our method demonstrates significantly higher qualities on both overall and top part of rankings then the state-of-art method and strong baselines.
Chinese Tweets Sentiment Analysis
I proposed a research project named ”Text Sentiment Analysis in SNS” and successfully received funding from China’s National Innovation Experiment Program for Undergraduates. To cope with the challenges introduced by the characteristics of text in SNS, I, under Prof. Zhoujun Li’s instruction, incorpo-rated several SNS elements into traditional sentiment analysis, including information dynamics in SNS, new words detection, and domain-dependent knowledge. Consequently, we used our model to participate in the China Computer Federation Conference on Natural Language Processing & Chinese Computing (NLP&CC 2012) Workshop for Weibo Sentiment Analysis Evaluation. Our system, among 34 teams and 53 submitted runs, ranked 1st in the task of Opinion Bearing Tweet Classification, and 4th in the task of Sentiment Polarity Classification.
Cloud Clip
Multi-device, especially multi-platform users suffer from transfering data and files among their different devices. While cloud storage and direct transmission are handy for large file synchronization, overhead would be too much if just small pieces of text or pictures were meant to be synced. Therefore, we developed a cloud clip synchronization tool to share clipboard for all platforms including Windows, Windows Phone, OS X, iOS, Linux, and Android.
Work Experiences
Microsoft Research Asia
Apr.2013 - Jan.2014 Research Intern
Mentored by Dr. Chin-Yew Lin and Researcher Yunbo Cao, I tried to express the intents about entity attributes by mining query intent templates from query log. The mining process consists of two steps: (a) Automatically collect various expressions on attributes and (b) Group and ground the expressions to the attributes defined in a knowledge base. We propose a framework that employs a weakly- supervised and metric learning integrated clustering method to construct query intent templates. The framework was integrated into Microsoft Bing’s query understanding module to improve the relevance and user experience of its searching results.
NExT Seach Center
Aug.2012 - Dec.2012 Part-time Developer
NExT is NUS-Tsinghua Extreme Search Centre (NExT), a leading research centre jointly established by the National University of Singapore (NUS) and Tsinghua University of China. I implemented a module of a distributed real-time crawler by analyzing the layouts and structures of user generated content websites like Twitter, Amazon, and Flicker.
MicroStrategy
Jun.2012 - Jul.2012 Intern Software Engineer
I developed and tested several modules of a web service system for MicroStrategy's Alert platform, using Java and Spring framework.