Machine Learning

• Harmonic mixtures: combining mixture models and graph-based methods for inductive and scalable semi-supervised learning. Xiaojin Zhu, John Lafferty. ICML 2005.
• Time-Sensitive Dirichlet Process Mixture Models. Xiaojin Zhu, Zoubin Ghahramani, John Lafferty. tech report CMU-CALD-05-104 , 2005.
• Person Identification in Webcam Images: An Application of Semi-Supervised Learning. Maria-Florina Balcan, Avrim Blum, Pakyan Choi, John Lafferty, Brian Pantano, Mugizi Robert Rwebangira and Xiaojin Zhu.  ICML2005 Workshop on Learning with Partially Classified Training Data, 2005.
• The FreeFoodCam dataset
  
• Semi-Supervised Learning with Graphs. Xiaojin Zhu. Doctoral thesis. CMU-LTI-05-192, May 2005.
  
• Semi-supervised learning literature survey, an online version of Chapter 11 that will keep updating.
• Java animation of random walk with absorbing states
• Nonparametric Transforms of Graph Kernels for Semi-Supervised Learning.   Xiaojin Zhu, Jaz Kandola, Zoubin Ghahramani, John Lafferty.   NIPS 2004.
  • Matlab code and datasets
  • Update 2005/4/10: the optimization can be formulated as Quadratic Programming instead of QCQP
• Kernel Conditional Random Fields: Representation and Clique Selection.  John Lafferty, Xiaojin Zhu, Yan Liu.  The Twenty-First International Conference on Machine Learning (ICML-2004). [ps | pdf]
• slides [slides pdf | slides ps.gz]
• tech report version: CMU-CS-04-115 [gzipped ps | pdf]
• notes on mean field training
• Semi-Supervised Learning: From Gaussian Fields to Gaussian Processes.   Xiaojin Zhu, John Lafferty, Zoubin Ghahramani.   CMU tech report CMU-CS-03-175 , 2003. [gzipped ps | pdf]
• Combining Active Learning and Semi-Supervised Learning Using Gaussian Fields and Harmonic Functions.   Xiaojin Zhu, John Lafferty, Zoubin Ghahramani.   ICML 2003 workshop on The Continuum from Labeled to Unlabeled Data in Machine Learning and Data Mining. [gzipped ps | pdf]
• Matlab code for combined active learning and semi-supervised learning
• Semi-Supervised Learning Using Gaussian Fields and Harmonic Functions.  Xiaojin Zhu, Zoubin Ghahramani, John Lafferty.  The Twentieth International Conference on Machine Learning (ICML-2003). [gzipped ps | pdf]
• slides [slides pdf | slides ps.gz]
• Matlab code for harmonic functions
• some semi-supervised learning datasets used in the ICML03 paper
• Learning from Labeled and Unlabeled Data with Label Propagation.  Xiaojin Zhu, Zoubin Ghahramani.  CMU CALD tech report CMU-CALD-02-107, 2002. [gzipped ps | pdf | short version]
• Towards Semi-Supervised Classification with Markov Random Fields.  Xiaojin Zhu, Zoubin Ghahramani.  CMU CALD tech report CMU-CALD-02-106, 2002. [pdf]

• Whole-Sentence Exponential Language Models: a Vehicle for Linguistic-Statistical Integration. Ronald Rosenfeld, Stanley F. Chen and Xiaojin Zhu.  Computers Speech and Language, 15(1), 2001.
• A Unified Design for Human-Machine Voice Interaction. Stefanie Shriver, Arthur Toth, Xiaojin Zhu, Alex Rudnicky, Roni Rosenfeld. Conference on Human Factors in Computing Systems (CHI) 2001.
• Improving Trigram Language Modeling with the World Wide Web. Xiaojin Zhu, Ronald Rosenfeld. International Conference on Acoustics, Speech and Signal Processing (ICASSP) 2001.[Also tech report CMU-CS-00-171]
• Towards a universal speech interface. Ronald Rosenfeld, Xiaojin Zhu, Arthur Toth, Stefanie Shriver, Kevin Lenzo, Alan W Black.  International Conference on Spoken Language Processing (ICSLP) 2000
• Linguistic Features For Whole Sentence Maximum Entropy Language Models.  Xiaojin Zhu,  Stanley F. Chen,  Ronald Rosenfeld.  5th European Conference on Speech Communication and Technology (Eurospeech), 1999 

 

• Segmenting Hands of Arbitrary Color. Xiaojin Zhu, Jie Yang, Alex Waibel. Fourth IEEE International Conference on Automatic Face and Gesture Recognition (F&G), 2000
• Multimodal People ID for Multimedia Meeting Browser. Jie Yang, Xiaojin Zhu, Ralph Gross, John Kominek, Yue Pan, Alex Waibel. The Seventh ACM International Multimedia Conference, 1999  

• Carnegie Mellon University, 2004/6/16, Learning from Labeled & Unlabeled Data, for CALD summer school Data Mining and Machine Learning: Algorithms and Applications
  
• Gatsby Computational Neuroscience Unit, 2004/3/10, ICML'04 reviewers behavior study [pdf | ps.gz]
  
• University of Cambridge Speech Research Group, 2004/3/2, Semi-supervised Learning with Gaussian Random Fields [pdf | ps.gz]
  
• Gatsby Computational Neuroscience Unit, 2004/2/27, Kernel Conditional Random Fields [slides pdf | slides ps.gz]
  
• Microsoft Research Cambridge, 2004/2/26, Semi-supervised Learning with Gaussian Random Fields [pdf | ps.gz]
  
• NSF Aladdin Workshop on Graph Partitioning in Vision and Machine Learning, 2003/1/10, Semi-Supervised Learning with Label Propagation
  

• If you are REALLY bored, take a look at some of my course project reports. Hear a talking clock in my voice!