Contact Information
Email: kchang10 AT illinois -DOT- edu
Kai-Wei ChangI am a Ph.D. student in University of Illinois at Urbana-Champaign supervised by Prof. Dan Roth. Before coming here, I received my master degree in National Taiwan University with a supervision of Prof. Chih-Jen Lin. My present research interests mainly focus on machine learning and nature language process. Please see my Curriculum Vitae for further information.Related Softwares & Research Projects:
|
|
Education & Experience
- Ph.D. student, Department of Computer Science, Aug. 2010 -- current.
- Member of the Cognitive Computation Group; Advisor: Prof. Dan Roth
- M.S., Department of Computer Science & Information Engineering, June 2009.
- GPA 4.00/4.00. Ranked 1st out of 158
- Member of the Machine Learning and Data Mining Group; Advisor: Prof. Chih-Jen Lin (Thesis)
- B.S., Department of Computer Science & Information Engineering, June 2007.
- B.S., Department of Electrical Engineering, (Dual degree), June 2007.
- GPA 4.00/4.00 (CS), 3.92/4.00 (EE), 3.96/4.00 (Total), Ranked 3rd out of 108, with 5 Presidential Awards (top 5% each semester)
- Member of the Machine Learning and Data Mining Group; Advisor: Prof. Chih-Jen Lin
Research Intern, Google Beijing Research, May 2008-Sep. 2008
We applied linear SVM solvers to the explicit form of polynomially mapped data and
improved the performance of a data-driven dependency parsing system. (see the
JMLR paper)
Awards
- Yahoo! Key Scientific Challenges Program Award, 2011
- Best Research Paper Award, KDD 2010
- Master Thesis Award, Taiwanese Association for Artificial Intelligence, 2009
Publications
- K.-W. Chang, D. Roth. Selective Block Minimization for Faster Convergence of Limited Memory Large-scale Linear Models. The 17th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining (KDD 2011) (Code,Poster,Slide)
- K.-W. Chang, R. Samdani, A. Rozovskaya, N. Rizzolo, M. Sammons, D. Roth. Inference Protocols for Coreference Resolution. Proceedings of the Fifteenth Conference on Computational Natural Language Learning: Shared Task (CoNLL 2011) (Our system ranked 3rd among 21 submissions) (Poster,Slide)
- H.-F. Yu, C.-J. Hsieh, K.-W. Chang, C.-J. Lin
Large linear classification when data cannot fit in memory.
ACM Transactions on Knowledge Discovery from Data Volumn 5 Issue 4 (2012).
(Code).
A preliminary version appeared in KDD 2010 and received (Best Research Paper Award) (KDD paper).
Another simplified version is published in IJCAI 2011 (Best paper track). (IJCAI paper) (Video)
Here is a 5-min introduction of this work (Video) - G.-X. Yuan, K.-W. Chang, C.-J. Hsieh, C.-J. Lin. A comparison of optimization methods for large scale L1-regularized linear classification.. Journal of Machine Learning Research (JMLR) 11(2010), 3183-3234. (Code)
- Y.-W. Chang, C.-J. Hsieh, K.-W. Chang, Michael Ringgaard, and C.-J. Lin. Training and Testing Low-degree Polynomial Data Mappings via Linear SVM. Journal of Machine Learning Research (JMLR) 11(2010), 1471-1490. (Code)
- F.-L. Huang, C.-J. Hsieh, K.-W. Chang, and C.-J. Lin. Iterative Scaling and Coordinate Descent Methods for Maximum Entropy Models. Journal of Machine Learning Research (JMLR) 11(2010), 815-848. A short version appears as a short paper at ACL 2009. (Code)
- H.-Y. Lo, K.-W. Chang, S.-T. Chen, T.-H. Chiang, C.-S. Ferng, C.-J. Hsieh, Y.-K. Ko, T.-T. Kuo, H.-C. Lai, K.-Y. Lin, C.-H. Wang, H.-F. Yu, C.-J. Lin, H.-T. Lin and S.-d. Lin. An Ensemble of Three Classifiers for KDD Cup 2009: Expanded Linear Model, Heterogeneous Boosting, and Selective Naive Bayes. JMLR Workshop and Conference Proceedings V.7, 57-64, 2009 (Third Place of the KDDCup'09 Slow Track).
- R.-E. Fan, K.-W. Chang, C.-J. Hsieh, X.-R. Wang, and C.-J. Lin. LIBLINEAR: A Library for Large Linear Classification. Journal of Machine Learning Research (JMLR) 9(2008), 1871-1874. (LIBLINEAR package)
- S. S. Keerthi, S. Sundararajan, K.-W. Chang, C.-J. Hsieh, and C.-J. Lin. A Sequential Dual Method for Large Scale Multi-ClassLinear SVMs. The 14th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining (KDD 2008).
- C.-J. Hsieh, K.-W. Chang, C.-J. Lin, S. S. Keerthi, and S. Sundararajan. A Dual Coordinate Descent Method for Large-scale Linear SVM. The 25th International Conference on Machine Learning (ICML 2008). (Code slide, video)
- K.-W. Chang, C.-J. Hsieh, and C.-J. Lin. Coordinate Descent Method for Large-scale L2-loss Linear SVM. Journal of Machine Learning Research (JMLR) 9(2008), 1369-1398. (Code)
Other Honors & Awards
- Honorary Member of the Phi Tau Phi Scholastic Honor Society, National Taiwan University, 2009
officially recommended by NTU, from top 3% of 156 graduating master students in CS department - Presidential Award, National Taiwan University, Fall 2003, Spring 2004, Fall 2004, Spring 2005, Fall 2006
given to the top 5% undergraduate students each semester - Third Prize in the Slow Track of KDDCUP 2009
out of more than 400 submissions - Winner in SVM track of Pascal Large Scale Learning Challenge in ICML 2008 Workshop
the only team that finished solving all huge problems - Student Travel Grant for ICML 2008, ICML
- Scholarship for Graduate Student, GARMIN, 2008
- Outstanding Students Conference Travel Grant, Foundation for The Advancement of Outstanding Scholarship, Taiwan, 2008
- Dean's List, National Taiwan University, 2006
for providing Calculus consulting - ACM ICPC Asia Regional programming Contest
Sixth Place (2003,2006), Seventh Place (2005), With Cho-Jui Hsieh, Peng-Ren Cheng - National College Programming Contest (Taiwan)
Outstanding Performance Award(2009), Second Prize (2005), with Cho-Jui Hsieh, Peng-Ren Cheng - Third Prize of physics, National High School Science Fair, 2002
with Cho-Jui Hsieh, Peng-Ren Cheng, and Li-Jen Chu