Xiaojin Zhu (Jerry) Assistant Professor Department of Computer Sciences University of Wisconsin-Madison 1210 West Dayton Street Madison, WI 53706-1685, USA Tel: 608-890-0129 Fax: 608-262-9777 Email: jerryzhu@cs.wisc.edu

Research Interests: machine learning (semi-supervised learning, latent variable models and topic models, Bayesian methods), and its applications in natural language processing, cognitive psychology, and human computer interaction.

Full publication list [sort by year | sort by topic ]

Recent Highlights:

• Semi-Supervised Learning
  • We published a book for beginners: Introduction to semi-supervised learning. Morgan & Claypool, 2009. [available online]
  • A concise, more technical summary of semi-supervised learning, in Encyclopedia of Machine Learning, to appear [pdf]
  • (AISTATS 09) Multi-manifold semi-supervised learning -- learning when data lives on multiple, intersecting manifolds. [pdf]
  • (NIPS 08) Unlabeled data: now it helps, now it doesn't -- a finite-sample minimax analysis on when semi-supervised learning is better than supervised learning. [pdf]
  • (ECML 08) Online semi-supervised learning -- learning when labeled and unlable items arrive sequentially. [pdf]
  • Semi-supervised learning tutorials [Chicago Summer School 2009, ICML 2007]
  • Semi-supervised learning literature survey [link]
• Latent Topic Models:
  • (ICML 09) Dirichlet forest priors -- shaping latent topic models with domain knowledge, coded as a mixture of Dirichlet trees [pdf]
  • (ECML 07) Statistical debugging using latent topic models [pdf]
• Computational Cognitive Psychology
  • (NIPS 09) Human Rademacher complexity -- bounding human generalization ability using computational learning theory [pdf]
  • (NIPS 08) Human active learning -- human learning rate is exponential (similar to active machine learning rate), if they can select queries [pdf]
  • (AAAI 08) Monkey online learning -- modeling concept-drift in Wisconsin Card Sort Task [pdf]
  • (AAAI 07) Human semi-supervised learning -- humans can learn from labeled and unlabeled data [pdf]
• Natural Language Processing
  • (NAACL 09) May all your wishes come true: A study of wishes and how to recognize them -- a new type of sentiment analysis [pdf]
  • (ACL 08) Learning bigrams from unigrams -- seemingly impossible, but the information is hidden in marginal distributions [pdf]
  • (AAAI 07) Text-to-Picture synthesis -- turning English sentences into pictures [pdf]
  • (NAACL 07) Improving diversity in ranking using absorbing random walks -- ranking items a la PageRank while maximizing diversity [pdf]

Current Professional Activities

• Editorial Board, Machine Learning Journal
• Program Committee / Referee: ICML 2010 (Area Chair), AAAI 2010, AISTATS 2010, COLING 2010, NIPS 2009, UAI 2009, ICML 2009, ACL 2009, NAACL 2009, AISTATS 2009
• North American Computational Linguistics Olympiad (NACLO 2010) University of Wisconsin-Madison site 2010
• AAAI 2009 Fall Symposium on Manifold Learning and its Applications
• University of Chicago Summer School/Workshop on Theory and Practice of Computational Learning, 2009
• NIPS 2008 Workshop on Machine Learning Meets Human Learning

Ph.D. Carnegie Mellon University, 2005. (C.V.)

• Fall 2010: (CS 540) Introduction to Artificial Intelligence
• Spring 2010: (CS 769) Advanced Natural Language Processing
• Spring 2010: Learning Math for Machine Learning (LMML) Reading Group (organized by David Andrzejewski)
• Fall 2009: (CS 540) Introduction to Artificial Intelligence
• Fall 2009: Nonparametric Bayesian Reading Group (organized by Andrew Goldberg)
• Fall 2009: HAMLET (Human, Animal, and Machine Learning: Experiment and Theory)
• AI at Wisconsin

Alumni
Lijie Heng, MS 2008, Oracle
Jurgen Van Gael, MS 2007, Ph.D. student at University of Cambridge