• General
• Education
• Projects
• Publications
• Honors
• Contact
• cv
• Others

General

I got my Bachelor degree from Fudan University . I joined Media Computing and Web Intelligence Lab led by Prof. Xiangyang Xue in Jan. 2006. My major research interest involves Natural Language Processing and Machine Learning. I am also interested in algorithm design. Now I'm moving to University of Wisconsin-Madison for graduate school. Please visit my new homepage.

Education

• Bachelor of Science

  Department of Computer Science and Engineering, Fudan University

  June 2008

• Preparatory student

  Department of Computer Science and Engineering, Fudan University

  Sep. 2004 till Sep. 2005

Projects

• Semi-Supervised Learning for Relation Extraction

  July 2007 till August 2007; UCLA IPAM RIPS Project 2007; MSRA

  This project aimed to find a new approach to do semi-supervised.Proposed a new method and compared it with Trasductive SVM provided by SVMLight and SDP. Tested on the ACE 2004 of Relation Extraction.

• Research on the Hierarchical Conditional Random Fields

  June. 2007 till now; National Undergraduate Innovational Experience Program; Fudan University

  This research focuses on developing Hierarchical Conditional Random Fields to address Chinese Part-Of-Speech Tagging on coarse documents (without pre-segmentation). Proposed a method to compute inference of HCRF efficiently under some assumptions. To compare the results with other approaches on Chinese POS Tagging.

  Midterm Report

• Research on Named Entity Recognition

  Dec. 2006 till May. 2007, Wangdao Research Project for Undergraduate; Fudan University

  Compare Maximum Entropy model with Conditional Random Fields model combining several local features and global features along with human knowledge to classify Chinese OOV words. Showed that human knowledge not only can improve the performance of the model, but also can smooth the model, and thus the training data might be reduced. Aslo demonstrated that global feature can overcome the problem of limited available local feature during the second or latter appearance of a Named Entity.

• Research on Word Sense Disambiguation

  Dec. 2005 till July. 2006, Media Computing and Web Intelligence Lab

  Used Weka to compare several popular classifiers and found that SVM performed best among all classifiers. Since the labeled training data was expensive to get, unsupervised learning is aslo valuable to investigate. Also showed that the use of WordNet to do an extended lesk can improve the performance of the system.

Publications

• Zhiting Xu. Bachelor Thesis (Chinese only). 2008

• Yuejie Zhang, Zhiting Xu, Tao Zhang. Fusion of Multiple Features for Chinese Named Entity Recognition based on CRF Model. Asia Information Retrieval Symposium (AIRS 2008), 2008

• Zhiting Xu, Xian Qian, Yuejie Zhang, Yaqian Zhou. CRF-based Hybrid Model for Word Segmentation, NER and even POS Tagging. SIGHAN Workshop on Chinese Language Processing 2008, 2008

• Yaofeng Wang, Yuejie Zhang, Zhiting Xu, Tao Zhang. Research On Dual Pattern Of Unsupervised And Supervised Word Sense Disambiguation. Proceedings of 2006 International Conference on Machine Learning and Cybernetics (ICMLC 2006), 2006

Honors

• Prize for Innovations in Research

• People’s Scholarship, 2nd Prize, Autumn, 2006, 2007

• WangDao Scholarship, Autumn, 2006

• 2005 ACM International Collegiate Programming Contest Regional ChengDu site 15th Place

• 2005 ACM International Collegiate Programming Contest Regional Beijing site 11th Place

• National Olympic Information Contest (NOI 2004), Bronze Medal

Contact

• Email : zhiting at fudan dot edu dot cn

  gtalk: zhiting dot xu at gmail dot com

  MSN: wshxzt at hotmail dot com

Others

• Related People

• I attended MLSS 2008, and here is my summary.

  
  Machine Learning Summer School 2008 - Kioloa
  

• I love jogging, and here is my log.

• I love Taekwondo, and now I am a Green Belt.

[ ... This website is underconstruction ... ]
last update: Aug 1, 2008 12:00 PM