Xiaojin (Jerry) Zhu

Professor

University of Wisconsin-Madison

2024

• Shubham Bharti, Stephen Wright, Adish Singla, and Xiaojin Zhu. On the complexity of teaching a family of linear behavior cloning learners. In Advances in Neural Information Processing Systems (NeurIPS), 2024.


• Jeremy McMahan, Young Wu, Yudong Chen, Xiaojin Zhu, and Qiaomin Xie. Inception: Efficiently computable misinformation attacks on Markov games. In The first Reinforcement Learning Conference (RLC), 2024.
  [arXiv | RLC]


• Young Wu, Jeremy McMahan, Yiding Chen, Yudong Chen, Xiaojin Zhu, and Qiaomin Xie. Minimally modifying a Markov game to achieve any Nash Equilibrium and value. In The 41st International Conference on Machine Learning (ICML), 2024.
  [arXiv version | ICML24 version pdf | code]


• Jeremy McMahan and Xiaojin Zhu. Anytime-constrained reinforcement learning. In The 27th International Conference on Artificial Intelligence and Statistics (AISTATS), 2024.
  [pdf]


• Young Wu, Jeremy McMahan, Xiaojin Zhu, and Qiaomin Xie. Data Poisoning to Fake a Nash Equilibrium in Markov Games. In The Thirty-Eighth AAAI Conference on Artificial Intelligence (AAAI), 2024.
  [pdf]


• Yiding Chen, Xuezhou Zhang, Qiaomin Xie, and Xiaojin Zhu. Exact policy recovery in offline RL with both heavy-tailed rewards and data corruption. In The Thirty-Eighth AAAI Conference on Artificial Intelligence (AAAI), 2024.
  [pdf]


• Jeremy McMahan, Young Wu, Xiaojin Zhu, and Qiaomin Xie. Optimal attack and defense on reinforcement learning. In The Thirty-Eighth AAAI Conference on Artificial Intelligence (AAAI), 2024.
  [pdf]


• Ara Vartanian, Xiaoxi Sun, Yun-Shiuan Chuang, Siddharth Suresh, Xiaojin Zhu, and Timothy Rogers. Learning interactions to boost human creativity with bandits and GPT-4. In The Annual Conference of the Cognitive Science Society (CogSci), 2024.
  [pdf]


• Jihyun Rho, Martina Rau, Shubham Bharti, Rosanne Luu, Jeremy McMahan, Andrew Wang, and Xiaojin Zhu. Various misleading visual features in misleading graphs: Do they truly deceive us? In The Annual Conference of the Cognitive Science Society (CogSci), 2024.
  [pdf | materials]


• Yun-Shiuan Chuang, Xiaojin Zhu, and Timothy Rogers. The delusional hedge algorithm as a model of human learning from diverse opinions. In The Annual Conference of the Cognitive Science Society (CogSci), 2024. Winner of the Computational Modeling Prize for Higher-Level Cognition for CogSci 2024.
  [pdf]


• Shubham Bharti, Shiyun Cheng, Jihyun Rho, Martina Rao, Xiaojin Zhu. CHARTOM: A Visual Theory-of-Mind Benchmark for Multimodal Large Language Models. ArXiv preprint.
  [arxiv]

2023

• Yiding Chen, Xiaojin Zhu, and Kirthevasan Kandasamy. Mechanism design for collaborative normal mean estimation. In Advances in Neural Information Processing Systems (NeurIPS), 2023.


• Xuefeng Du, Yiyou Sun, Xiaojin Zhu, and Yixuan Li. Dream the impossible: Outlier imagination with diffusion models. In Advances in Neural Information Processing Systems (NeurIPS), 2023.


• Yiding Chen, Xuezhou Zhang, Kaiqing Zhang, Mengdi Wang, and Xiaojin Zhu. Byzantine robust online and offline distributed reinforcement learning. In The 26th International Conference on Artificial Intelligence and Statistics (AISTATS), 2023.
  [pdf]


• Leitian Tao, Xuefeng Du, Xiaojin Zhu, and Yixuan Li. Non-parametric outlier synthesis. In The 11th International Conference on Learning Representations (ICLR), 2023.


• Young Wu, Jeremy McMahan, Xiaojin Zhu, and Qiaomin Xie. Reward poisoning attacks on offline multi-agent reinforcement learning. In The Thirty-Seventh AAAI Conference on Artificial Intelligence (AAAI), 2023.
  [pdf | slides]
  Install a fake Dominant Strategy Equilibrium in offline batch MARL data, cheap!

2022

• Shubham Bharti, Xuezhou Zhang, Adish Singla, and Xiaojin Zhu. Provable defense against backdoor policies in reinforcement learning. In Advances in Neural Information Processing Systems (NeurIPS), 2022.
  [pdf | demo video (8MB MP4) | github code]
  Allows safe usage of a backdoored RL policy, without retraining.


• Yiyou Sun, Yifei Ming, Xiaojin Zhu, and Yixuan Li. Out-of-distribution detection with deep nearest neighbors. In The 39th International Conference on Machine Learning (ICML), 2022.


• Yuzhe Ma, Young Wu, and Xiaojin Zhu. Game redesign in no-regret game playing. In The 31st International Joint Conference on Artificial Intelligence and the 25th European Conference on Artificial Intelligence (IJCAI-ECAI 22), 2022.
  [pdf | short talk]
  Install a fake Dominant Strategy Equilibrium in two-player general-sum games, cheap!


• Xuezhou Zhang, Yiding Chen, Xiaojin Zhu, and Wen Sun. Corruption-robust offline reinforcement learning. In The 25th International Conference on Artificial Intelligence and Statistics (AISTATS). 2022
  [pdf]
  Even when offline data is epsilon-corrupted, one can learn a good RL policy from it.

2021

• Xuezhou Zhang, Yiding Chen, Xiaojin Zhu, and Wen Sun. Robust policy gradient against strong data corruption. In The 38th International Conference on Machine Learning (ICML). 2021
  [pdf | code and videos]
  Even when online RL experience is epsilon-corrupted, one can learn a good policy.


• Xuezhou Zhang, Shubham Bharti, Yuzhe Ma, Adish Singla, and Xiaojin Zhu. The sample complexity of teaching by reinforcement on Q-learning. In The Thirty-Fifth AAAI Conference on Artificial Intelligence (AAAI), 2021
  [pdf | poster | slides]


• Yuzhe Ma, Jon Sharp, Ruizhe Wang, Earlence Fernandes, and Xiaojin Zhu. Adversarial attacks on Kalman filter-based forward collision warning systems. In The Thirty-Fifth AAAI Conference on Artificial Intelligence (AAAI), 2021
  [pdf | poster]


• Yun-Shiuan Chuang, Xuezhou Zhang, Yuzhe Ma, Mark Ho, Joe Austerweil, and Xiaojin Zhu. Using machine teaching to investigate human assumptions when teaching reinforcement learners. In The 41st Annual Conference of the Cognitive Science Society (CogSci). 2021.


• Claudia Ramly, Ayon Sen, Ved P. Kale, Martina A. Rau, and Xiaojin Zhu. Digitally training graph viewers against misleading bar charts. In The 41st Annual Conference of the Cognitive Science Society (CogSci). 2021.
  [pdf]

2020

• Xuezhou Zhang, Yuzhe Ma, Adish Singla, and Xiaojin Zhu. Adaptive reward-poisoning attacks against reinforcement learning. In The 37th International Conference on Machine Learning (ICML). 2020.
  [pdf | arXiv | code]


• Amin Rakhsha, Goran Radanovic, Rati Devidze, Xiaojin Zhu, and Adish Singla. Policy teaching via environment poisoning: Training-time adversarial attacks against reinforcement learning. In The 37th International Conference on Machine Learning (ICML). 2020.
  [arXiv]


• Xuezhou Zhang, Xiaojin Zhu, and Laurent Lessard. Online Data Poisoning Attacks. In Learning for Dynamics and Control (L4DC). 2020.
  [arXiv]


• Yiding Chen and Xiaojin Zhu. Optimal attack against autoregressive models by manipulating the environment. In The Thirty-Fourth AAAI Conference on Artificial Intelligence (AAAI), 2020.
  [pdf | arXiv | matlab code]


• Ayon Sen, Xiaojin Zhu, Erin Marshall, Robert Nowak. Popular Imperceptibility Measures in Visual Adversarial Attacks are Far from Human Perception. In Conference on Decision and Game Theory for Security (GameSec), 2020.
  [pdf]

2019

• Yuzhe Ma, Xuezhou Zhang, Wen Sun, and Xiaojin Zhu. Policy poisoning in batch reinforcement learning and control. In Advances in Neural Information Processing Systems (NeurIPS), 2019.
  [pdf | poster | code | arXiv]


• Xuanqing Liu, Si Si, Xiaojin Zhu, Yang Li, and Cho-Jui Hsieh. A unified framework for data poisoning attack to graph-based semi-supervised learning. In Advances in Neural Information Processing Systems (NeurIPS), 2019.


• Farnam Mansouri, Yuxin Chen, Ara Vartanian, Xiaojin Zhu, and Adish Singla. Preference-based batch and sequential teaching: Towards a unified view of models. In Advances in Neural Information Processing Systems (NeurIPS), 2019.
  [arXiv]


• Ayon Sen, Xiaojin Zhu, Liam Marshall, Robert Nowak. Should Adversarial Attacks Use Pixel p-Norm?. arXiv:1906.02439. 2019.
  [link | code and data ]


• Yuzhe Ma, Xiaojin Zhu, Justin Hsu. Data Poisoning against Differentially-Private Learners: Attacks and Defenses. In The 28th International Joint Conference on Artificial Intelligence (IJCAI), 2019.
  [pdf | arxiv | github code]


• Sanjoy Dasgupta, Daniel Hsu, Stefanos Poulis, Xiaojin Zhu. Teaching a black-box learner. In The 36th International Conference on Machine Learning (ICML), 2019.
  [pdf]


• Laurent Lessard, Xuezhou Zhang, and Xiaojin Zhu. An optimal control approach to sequential machine teaching. In The 22nd International Conference on Artificial Intelligence and Statistics (AISTATS), 2019.
  [pdf | arXiv 1810.06175 | code]


• Owen Levin, Zihang Meng, Vikas Singh, Xiaojin Zhu. Fooling Computer Vision into Inferring the Wrong Body Mass Index. arXiv:1905.06916, 2019.
  [link]

2018

• Xiaojin Zhu. An optimal control view of adversarial machine learning. arXiv:1811.04422, 2018.
  [link | pdf]


• Kwang-Sung Jun, Lihong Li, Yuzhe Ma, and Xiaojin Zhu. Adversarial attacks on stochastic bandits. In Advances in Neural Information Processing Systems (NeurIPS), 2018.
  [pdf]


• Xiaojin Zhu, Adish Singla, Sandra Zilles, Anna N. Rafferty. An Overview of Machine Teaching. ArXiv 1801.05927, 2018.


• Yuzhe Ma, Robert Nowak, Philippe Rigollet, Xuezhou Zhang, and Xiaojin Zhu. Teacher improves learning by selecting a training subset. In The 21st International Conference on Artificial Intelligence and Statistics (AISTATS), 2018.
  [pdf | AMPL code]


• Xuezhou Zhang, Xiaojin Zhu, and Stephen Wright. Training set debugging using trusted items. In The Thirty-Second AAAI Conference on Artificial Intelligence (AAAI), 2018
  [pdf | code]


• Ayon Sen, Scott Alfeld, Xuezhou Zhang, Ara Vartanian, Yuzhe Ma, and Xiaojin Zhu. Training set camouflage. In Conference on Decision and Game Theory for Security (GameSec), 2018
  [pdf | code]


• Yuzhe Ma, Kwang-Sung Jun, Lihong Li, and Xiaojin Zhu. Data poisoning attacks in contextual bandits. In Conference on Decision and Game Theory for Security (GameSec), 2018
  [arXiv]


• Vraj Shah, Arun Kumar, and Xiaojin Zhu. Are key-foreign key joins safe to avoid when learning highcapacity classifiers?. In VLDB, 2018.


• Evan Hernandez, Ara Vartanian, and Xiaojin Zhu. Program synthesis with visual specification. ArXiv 1806.00938, 2018.
  [arXiv | visual specification corpus data set | speech programming game for Chrome | HAMLET talk | CS NEST talk | whitepaper]


• Ayon Sen, Purav Patel, Martina A. Rau, Blake Mason, Robert Nowak, Timothy T. Rogers, and Xiaojin Zhu.
  • For teaching perceptual fluency, machines beat human experts. In The 40th Annual Conference of the Cognitive Science Society (CogSci), 2018. [pdf]
  • Machine beats human at sequencing visuals for perceptual-fluency practice. In Educational Data Mining, 2018. [pdf]


• Robert M. Nosofsky, Craig A. Sanders, Xiaojin Zhu, and Mark A. McDaniel. Model-guided search for optimal natural-science-category training exemplars: A work in progress. Psychonomic Bulletin & Review, 2018.

2017

• Xiaojin Zhu, Ji Liu, and Manuel Lopes. No learner left behind: On the complexity of teaching multiple learners simultaneously. In The 26th International Joint Conference on Artificial Intelligence (IJCAI), 2017.
  Minimax teaching dimension to make the worst learner in a class learn. Partitioning the class into sections improves teaching dimension.
  [pdf]


• Scott Alfeld, Xiaojin Zhu, and Paul Barford. Explicit defense actions against test-set attacks. In The Thirty-First AAAI Conference on Artificial Intelligence (AAAI), 2017.
  [pdf]


• Paul Bennett, David M. Chickering, Christopher Meek, and Xiaojin Zhu. Algorithms for active classifier selection: maximizing recall with precision constraints. In The Tenth ACM International Conference on Web Search and Data Mining (WSDM), 2017.
  [pdf]

2016

• Ji Liu and Xiaojin Zhu. The teaching dimension of linear learners. Journal of Machine Learning Research, 17(162):1-25, 2016.
  This is the journal version of the ICML'16 paper, with a discussion on teacher-learner collusion.
  [link]


• Tzu-Kuo Huang, Lihong Li, Ara Vartanian, Saleema Amershi, and Xiaojin Zhu. Active learning with oracle epiphany. In Advances in Neural Information Processing Systems (NIPS), 2016.
  This paper brings active learning theory and practice closer. We analyze active learning query complexity where the oracle initially may not know how to answer queries from a certain region in the input space. After seeing multiple queries from the region, the oracle can have an "epiphany", i.e. realizing how to answer any queries from that region.
  [pdf]


• Ji Liu, Xiaojin Zhu, and H. Gorune Ohannessian. The Teaching Dimension of Linear Learners. In The 33rd International Conference on Machine Learning (ICML), 2016.
  We provide lower bounds on training set size to perfectly teach a linear learning. We also provide the corresponding upper bounds (and thus teaching dimension) by exhibiting teaching sets for SVM, logistic regression, and ridge regression.
  [pdf | supplementary | arXiv preprint]


• Jina Suh, Xiaojin Zhu, and Saleema Amershi. The label complexity of mixed-initiative classifier training. In The 33rd International Conference on Machine Learning (ICML), 2016.
  Do you do interactive machine learning with a human oracle? Then don't use active learning alone: mixing it with machine teaching is far better both in theory and in practice.
  [pdf | supplementary | slides]


• Xiaojin Zhu, Ara Vartanian, Manish Bansal, Duy Nguyen, and Luke Brandl. Stochastic multiresolution persistent homology kernel. In The 25th International Joint Conference on Artificial Intelligence (IJCAI), 2016.
  A kernel built upon persistent homology at multiple resolutions, and with Monte Carlo to speed up. Ready to use as topological features for machine learning.
  [pdf]


• Kwang-Sung Jun, Kevin Jamieson, Rob Nowak, and Xiaojin Zhu. Top arm identification in multi-armed bandits with batch arm pulls. In The 19th International Conference on Artificial Intelligence and Statistics (AISTATS), 2016.
  [pdf]


• Scott Alfeld, Xiaojin Zhu, and Paul Barford. Data Poisoning Attacks against Autoregressive Models. In The Thirtieth AAAI Conference on Artificial Intelligence (AAAI), 2016.
  Machine teaching for autoregression, applied to computer security.
  [pdf]


• Arun Kumar, Jeffrey Naughton, Jignesh M. Patel, and Xiaojin Zhu. To join or not to join? thinking twice about joins before feature selection. In ACM SIGMOD, 2016.
  


• Gabriel Cadamuro, Ran Gilad-Bachrach, and Xiaojin Zhu. Debugging machine learning models. In ICML Workshop on Reliable Machine Learning in the Wild, 2016.
  Training data repair to ensure certain test items are correctly predicted. An application of machine teaching.
  [pdf | extended abstract for CHI 2016 workshop on human centred machine learning]


• Felice Resnik, Amy Bellmore, Junming Xu, and Xiaojin Zhu. Celebrities emerge as advocates in tweets about bullying. In Translational Issues in Psychological Science, 2016.
  [pdf]


• Anna Kaatz, You-Geon Lee, Aaron Potvien, Wairimu Magua, Amarette Filut, Anupama Bhattacharya, Renee Leatherberry, Xiaojin Zhu, and Molly Carnes. Analysis of National Institutes of Health R01 application critiques, impact, and criteria scores: Does the sex of the principal investigator make a difference? In Academic medicine: journal of the Association of American Medical Colleges, 91(8):1080-8, 2016.
  


• Christopher Meek, Patrice Y. Simard, and Xiaojin Zhu. Analysis of a design pattern for teaching with features and labels, 2016. arXiv


• Xuezhou Zhang, Hrag Gorune Ohannessian, Ayon Sen, Scott Alfeld and Xiaojin Zhu. Optimal Teaching for Online Perceptrons. In NIPS 2016 workshop on Constructive Machine Learning, 2016. [pdf]

2015

• Kwang-Sung Jun, Xiaojin Zhu, Timothy Rogers, Zhuoran Yang, and Ming Yuan. Human memory search as initial-visit emitting random walk. In Advances in Neural Information Processing Systems (NIPS), 2015.
  A random walk that only emits an output when it visits a state for the first time.
  [pdf | supplemental | poster]


• Gautam Dasarathy, Robert Nowak, and Xiaojin Zhu. S²: An efficient graph based active learning algorithm with application to nonparametric classification. In Conference on Learning Theory (COLT), 2015.
  Active learning on graphs, with guarantees.
  [pdf | poster | slides | arxiv version]


• Shike Mei and Xiaojin Zhu. The security of latent Dirichlet allocation. In The Eighteenth International Conference on Artificial Intelligence and Statistics (AISTATS), 2015.
  How might an attacker poison the corpus to manipulate LDA topics? We answer this question via machine teaching.
  [pdf | slides ]


• Xiaojin Zhu. Machine Teaching: an Inverse Problem to Machine Learning and an Approach Toward Optimal Education. In The Twenty-Ninth AAAI Conference on Artificial Intelligence (Senior Member Track, AAAI), 2015. AAAI / Computing Community Consortium "Blue Sky Ideas" Track Prize.
  An overview of machine teaching.
  [pdf | talk slides | project link]


• Shike Mei and Xiaojin Zhu. Using Machine Teaching to Identify Optimal Training-Set Attacks on Machine Learners. In The Twenty-Ninth AAAI Conference on Artificial Intelligence (AAAI), 2015.
  An application of machine teaching to identify the optimal training-set-attacks against a learning algorithm.
  [pdf | poster ad | poster | Mendota ice data | Tech Report 1813]


• Shike Mei and Xiaojin Zhu. Some Submodular Data-Poisoning Attacks on Machine Learners. Computer Science Tech Report 1822, University of Wisconsin-Madison, 2015.
  [pdf]


• Bryan Gibson, Timothy Rogers, Charles Kalish, and Xiaojin Zhu. What causes category-shifting in human semi-supervised learning? In The 32nd Annual Conference of the Cognitive Science Society (CogSci), 2015.
  [pdf]


• Newsha Ardalani, Clint Lestourgeon, Karthikeyan Sankaralingam, and Xiaojin Zhu. Cross-architecture performance prediction (XAPP) using CPU code to predict GPU performance. In Annual IEEE/ACM International Symposium on Microarchitecture (MICRO-48), 2015.

2014

• Kaustubh Patil, Xiaojin Zhu, Lukasz Kopec, and Bradley Love. Optimal Teaching for Limited-Capacity Human Learners. In Advances in Neural Information Processing Systems (NIPS), 2014. Spotlight presentation.
  Using machine teaching we construct an optimal training data set to teach human students a categorization task, assuming the students use GCM as the learning algorithm. Our optimal training data set is non-iid, has interesting "idealization" properties, and outperforms iid training data sets sampled from the underlying test distribution.
  [pdf | poster | spotlight | data]


• Shike Mei, Jun Zhu, and Xiaojin Zhu. Robust RegBayes: Selectively incorporating First-Order Logic domain knowledge into Bayesian models. In the 31st International Conference on Machine Learning (ICML), 2014.
  [pdf | slides | poster ]


• Shike Mei, Han Li, Jing Fan, Xiaojin Zhu, and Charles R. Dyer. Inferring air pollution by sniffing social media. In the 2014 IEEE/ACM International Conference on Advances in Social Network Analysis and Mining (ASONAM), 2014.
  We can estimate real-time air quality from Weibo text posts. This can be useful in places where there is no physical air monitoring stations.
  [pdf | poster | data (WeiboAQIv1.tgz, 370MB) | news article]


• Amy Bellmore, Angela Calvin, Jun-Ming Xu, and Xiaojin Zhu. The five W's of bullying on Twitter: Who, what, why, where, when. In Computers in Human Behavior, 2014. Accepted.
  


• Angela J. Calvin, Amy Bellmore, Jun-Ming Xu, and Xiaojin Zhu. #bully: Uses of Hashtags in Posts about Bullying on Twitter. In Journal of School Violence, 2014. Accepted.
  


• Jun-Ming Xu, Hsun-Chih Huang, Amy Bellmore, and Xiaojin Zhu. School Bullying in Twitter and Weibo: a Comparative Study. In the Eighth International AAAI Conference on Weblogs and Social Media (ICWSM), 2014.
  [pdf]


• Charles Kalish, Xiaojin Zhu, and Timothy Rogers. Drift in children's categories: When experienced distributions conflict with prior learning. In Developmental Science, 2014.
  


• C. Gokhale, S. Das, A. Doan, J. Naughton, R. Rampalli, J. Shavlik, X. Zhu. Corleone: Hands-off Crowdsourcing for Entity Matching. In SIGMOD, 2014.
  


• Mark Liu, Mutlu Ozdogan, and Xiaojin Zhu. Crop type classification by simultaneous use of satellite images of different resolutions. In IEEE Transactions on Geoscience and Remote Sensing, 52(6):3637-3649, June 2014.

2013

• Xiaojin Zhu. Machine teaching for Bayesian learners in the exponential family. In Advances in Neural Information Processing Systems (NIPS), 2013.
  We study machine teaching, or optimal teaching, the inverse problem of machine learning.
  [pdf | poster | code]


• Kwang-Sung Jun, Xiaojin Zhu, Burr Settles, and Timothy Rogers. Learning from Human-Generated Lists. In The 30th International Conference on Machine Learning (ICML), 2013.
  Quick! Say as many animals as you can think of in 60 seconds without repetition! The list you produce is non-iid, non-exchangeable, and carries important information about animals (and your brain).
  [pdf | slides | SWIRL v1.0 code | video]


• Xiaojin Zhu. Persistent homology: An introduction and a new text representation for natural language processing. In The 23rd International Joint Conference on Artificial Intelligence (IJCAI), 2013.
  A gentle tutorial on homology, and an application in machine learning.
  [pdf | slides (long, short) | poster | data and code ]


• Jun-Ming Xu, Aniruddha Bhargava, Robert Nowak, and Xiaojin Zhu. Socioscope: Spatio-temporal signal recovery from social media (extended abstract). In The 23rd International Joint Conference on Artificial Intelligence (IJCAI), 2013.
  This is an invited abstract of our ECML/PKDD 2012 paper of the same title. More concise and readable.
  [pdf | code]


• Junming Xu, Benjamin Burchfiel, Xiaojin Zhu, and Amy Bellmore. An examination of regret in bullying tweets. In North American Chapter of the Association for Computational Linguistics - Human Language Technologies (NAACL HLT), short paper, 2013.
  4% users delete their bullying tweets. Why?
  [pdf | slides]


• Nick Bridle and Xiaojin Zhu. p-voltages: Laplacian regularization for semi-supervised learning on high-dimensional data. In Eleventh Workshop on Mining and Learning with Graphs (MLG2013), 2013.
  [pdf (errata) | slides | poster | code]


• Yimin Tan and Xiaojin Zhu. Dragging: Density-ratio bagging. Technical Report Computer Science TR1795, University of Wisconsin-Madison, 2013.
  [pdf]


• Bryan R. Gibson, Timothy T. Rogers, and Xiaojin Zhu. Human semi-supervised learning. Topics in Cognitive Science, 5(1):132-172, 2013.
  [link | data]

2012

• Jun-Ming Xu, Aniruddha Bhargava, Robert Nowak, and Xiaojin Zhu. Socioscope: Spatio-temporal signal recovery from social media. In The European Conference on Machine Learning and Principles and Practice of Knowledge Discovery in Databases (ECML PKDD), 2012. Best paper in knowledge discovery.
  Accurately estimating the underlying intensity function for any target phenomenon from noisy social media data.
  [pdf | code | slides | poster]


• Jun-Ming Xu, Kwang-Sung Jun, Xiaojin Zhu, and Amy Bellmore. Learning from bullying traces in social media. In North American Chapter of the Association for Computational Linguistics - Human Language Technologies (NAACL HLT). 2012.
  Let's fight bullying with machine learning! We show how several research problems in the study of bullying can be formulated as familiar natural language processing tasks. We also share our data and code. [pdf | data and code |slides | news article]


• Burr Settles and Xiaojin Zhu. Behavioral factors in interactive training of text classifiers. In North American Chapter of the Association for Computational Linguistics - Human Language Technologies (NAACL HLT). Short paper. 2012.
  [pdf]


• Jun-Ming Xu, Xiaojin Zhu, and Amy Bellmore. Fast learning for sentiment analysis on bullying. In ACM KDD Workshop on Issues of Sentiment Discovery and Opinion Mining (WISDOM), 2012.
  [pdf | slides]

2011

• Faisal Khan, Xiaojin Zhu, and Bilge Mutlu. How do humans teach: On curriculum learning and teaching dimension. In Advances in Neural Information Processing Systems (NIPS) 25. 2011.
  What is the optimal teaching strategy for a threshold function in 1D? Should one start teaching with items around the threshold? Or with the most unambiguous items farthest from the threshold? Two computational theories, teaching dimension and curriculum learning, disagree. We show that humans do the latter. We then extend teaching dimension theory to explain it.
  [pdf | data | slides | UCSD teaching workshop talk]


• Shilin Ding, Grace Wahba, and Xiaojin Zhu. Learning higher-order graph structure with features by structure penalty. In Advances in Neural Information Processing Systems (NIPS) 25. 2011.
  [pdf]


• Jun-Ming Xu, Xiaojin Zhu, and Timothy T. Rogers. Metric learning for estimating psychological similarities. ACM Transactions on Intelligent Systems and Technology (ACM TIST), 2011.
  [journal link | unofficial version | data | code]


• David Andrzejewski, Xiaojin Zhu, Mark Craven, and Ben Recht. A framework for incorporating general domain knowledge into Latent Dirichlet Allocation using First-Order Logic. The Twenty-Second International Joint Conference on Artificial Intelligence (IJCAI-11), 2011.
  Want to add all kinds of domain knowledge (e.g., word-must-in-this-topic, word-must-not-in-that-topic, if-this-then-that, etc.) to LDA? As long as you can write them in FOL, our fold.all model can combine your logic knowledge base with latent Dirichlet allocation for you via stochastic gradient descent. The resulting topics will be guided by both logic and data statistics. You don't have to derive customized LDA variants ever again (disclaimer: read the paper). In other words, fold.all to LDA is like constrained clustering to clustering. [pdf | slides | poster | data (68MB) | code]


• Xiaojin Zhu, Bryan Gibson, and Timothy Rogers. Co-training as a human collaboration policy. In The Twenty-Fifth Conference on Artificial Intelligence (AAAI-11), 2011.
  We turn the co-training algorithm into a human collaboration policy, where two people together learn a concept. Importantly, each person sees half of the features, and the two communicate by exchanging labels. We show that this policy leads to unique nonlinear decision boundaries that were difficult to learn had the two people fully collaborated or did not collaborate. [pdf | data | slides]


• Andrew Goldberg, Xiaojin Zhu, Alex Furger, and Jun-Ming Xu. OASIS: Online active semisupervised learning. In The Twenty-Fifth Conference on Artificial Intelligence (AAAI-11), 2011.
  A Bayesian model to learn a classifier from streaming data with missing labels. It also asks active learning questions. [pdf | slides]


• Chen Yu, Jun-Ming Xu, and Xiaojin Zhu. Word learning through sensorimotor child-parent interaction: A feature selection approach. The 33rd Annual Conference of the Cognitive Science Society (CogSci 2011), 2011.
  [pdf]


• Charles W. Kalish, Timothy T. Rogers, Jonathan Lang, and Xiaojin Zhu. Can semi-supervised learning explain incorrect beliefs about categories? Cognition, 2011.
  Another study of test-item effects (see our ICML'10 paper) in humans, explained by semi-supervised learning models. [link]


• Arthur Glenberg, Jonathan Willford, Bryan Gibson, Andrew Goldberg, and Xiaojin Zhu. Improving reading to improve math. Scientific Studies in Reading, 2011.
  [pdf]


• Nathan Rosenblum, Xiaojin Zhu, and Barton P. Miller. Who wrote this code? identifying the authors of program binaries. In The European Symposium on Research in Computer Security (ESORICS), 2011.
  [pdf]


• Nathan Rosenblum, Barton P. Miller, and Xiaojin Zhu. Recovering the toolchain provenance of binary code. In International Symposium on Software Testing and Analysis (ISSTA), 2011. ACM SIGSOFT Distinguished Paper Award.
  [pdf]


• Mariyam Mirza, Paul Barford, Xiaojin Zhu, Suman Banerjee, and Michael Blodgett. Fingerprinting 802.11 rate adaptation algorithms. In The 30th IEEE International Conference on Computer Communications (INFOCOM), Shanghai, China, 2011.
  [pdf]


• Jake Rosin, Andrew B. Goldberg, Xiaojin Zhu, and Charles Dyer. A Bayesian model for image sense ambiguity in pictorial communication systems. In Technical Report Computer Science TR1692, University of Wisconsin-Madison, 2011.
  [pdf]


• Xiaojin Zhu, Jun-Ming Xu, Christine M. Marsh, Megan K. Hines, and F. Joshua Dein. Machine learning for zoonotic emerging disease detection. In ICML 2011 Workshop on Machine Learning for Global Challenges, 2011.
  West Nile virus! We might have detected it earlier ten years ago, had people reported that they were seeing dead crows in their backyards. This position paper suggests a machine learning system for far upstream detection of zoonotic disease outbreaks. [pdf | poster]

2010

• Bryan Gibson, Xiaojin Zhu, Tim Rogers, Chuck Kalish, and Joseph Harrison. Humans learn using manifolds, reluctantly. In Advances in Neural Information Processing Systems (NIPS) 24, 2010. Oral presentation.
  Humans can learn the two-moon dataset, if we give them 4 (but not 2) labeled points and clue them in on the graph. [pdf | NIPS talk slides]


• Andrew Goldberg, Xiaojin Zhu, Benjamin Recht, Jun-Ming Xu, and Robert Nowak. Transduction with matrix completion: Three birds with one stone. In Advances in Neural Information Processing Systems (NIPS) 24. 2010.
  Find it difficult to do transductive learning on multi-label data with many missing features and missing labels? Let matrix completion help. [pdf | poster]


• Xiaojin Zhu, Bryan R. Gibson, Kwang-Sung Jun, Timothy T. Rogers, Joseph Harrison, and Chuck Kalish. Cognitive models of test-item effects in human category learning. In The 27th International Conference on Machine Learning (ICML), 2010.
  Two people with exactly the same training may classify a test item differently, depending on what other test items they are asked to classify (without label feedback). We explain such Test-Item Effect with online semi-supervised learning, which extends the exemplar, the prototype and the rational models of categorization. [paper pdf | slides | poster]


• Bryan R Gibson, Kwang-Sung Jun, and Xiaojin Zhu. With a little help from the computer: Hybrid human-machine systems on bandit problems. In NIPS 2010 Workshop on Computational Social Science and the Wisdom of Crowds, 2010.
  [pdf | slides at NIPS'09 workshop]


• Faisal Khan, Bilge Mutlu, and Xiaojin Zhu. Modeling social cues: Effective features for predicting listener nods. In NIPS 2010 Workshop on Human Communication Dynamics, 2010.
  [pdf | slides]


• Timothy Rogers, Charles Kalish, Bryan Gibson, Joseph Harrison, and Xiaojin Zhu. Semi-supervised learning is observed in a speeded but not an unspeeded 2D categorization task. In Proceedings of the 32nd Annual Conference of the Cognitive Science Society (CogSci), 2010.
  [paper pdf]


• Nathan Rosenblum, Barton Miller, and Xiaojin Zhu. Extracting compiler provenance from program binaries. In Proceedings of the 9th ACM SIGPLAN-SIGSOFT workshop on Program Analysis for Software Tools and Engineering (PASTE), 2010. [paper pdf]


• David Andrzejewski, David G. Stork, Xiaojin Zhu, and Ron Spronk. Inferring compositional style in the neo-plastic paintings of Piet Mondrian by machine learning. In Electronic Imaging: Computer Image Analysis in the Study of Art (SPIE 2010), 2010. [paper pdf | slides ppt | data]

2009

• Xiaojin Zhu, Timothy Rogers, and Bryan Gibson. Human Rademacher Complexity. In Advances in Neural Information Processing Systems (NIPS) 23, 2009.
  This student in your class keeps nodding and can recite everything you said. How do you know if he has truly learned the material, or is he simply overfitting your lecture? We offer a measure that combines computational learning theory and cognitive psychology to gauge human generalization abilities.
  [paper pdf | poster | the Shape domain images | the Word domain text | task=WordLength example subject file]


• Xiaojin Zhu and Andrew B. Goldberg. Introduction to Semi-Supervised Learning. Morgan & Claypool, 2009.
  A short, self-contained introductory book to semi-supervised learning. For advanced undergraduates, entry-level graduate students and researchers in Computer Science, Electrical Engineering, Statistics, Psychology, etc. You may already have access to the book through your institution -- check "Access" on the link. [link]


• Xiaojin Zhu. Semi-Supervised Learning. Encyclopedia entry in Claude Sammut and Geoffrey Webb, editors, Encyclopedia of Machine Learning. Springer, to appear.
  A concise, technical summary of semi-supervised learning. [pdf]


• David Andrzejewski, Xiaojin Zhu, and Mark Craven. Incorporating domain knowledge into topic modeling via Dirichlet forest priors. In The 26th International Conference on Machine Learning (ICML), 2009.
  Allowing Must-links and Cannot-links on words in LDA topics, by replacing the Dirichlet prior with a mixture of Dirichlet trees. [pdf | slides | poster | code]


• Andrew Goldberg, Xiaojin Zhu, Aarti Singh, Zhiting Xu, and Robert Nowak. Multi-manifold semi-supervised learning. In Twelfth International Conference on Artificial Intelligence and Statistics (AISTATS), 2009.
  What if the data consists of multiple, intersecting manifolds? We cut the ambient space into pieces by clustering unlabeled data using a Hellinger distance metric sensitive to manifold dimensionality, orientation, and density. Within each piece, we then perform supervised learning using the labeled data. [pdf | poster | dataset]


• Andrew Goldberg, Nathanael Fillmore, David Andrzejewski, Zhiting Xu, Bryan Gibson, and Xiaojin Zhu. May all your wishes come true: A study of wishes and how to recognize them. In North American Chapter of the Association for Computational Linguistics - Human Language Technologies (NAACL HLT), 2009.
  People from around the world offered up their wishes to be printed on confetti and dropped from the sky during the famous New Year's Eve ``ball drop'' in New York City's Times Square. We present an in-depth analysis of this collection of wishes. We then leverage this unique resource to conduct the first study on building general ``wish detectors'' for natural language text. [pdf | slides | data]


• Arthur Glenberg, Andrew B. Goldberg, and Xiaojin Zhu. Improving early reading comprehension using embodied CAI. Instructional Science, 2009.
  [pdf]


• Andrew B. Goldberg and Xiaojin Zhu. Keepin' it real: Semi-supervised learning with realistic tuning. In NAACL 2009 Workshop on Semi-supervised Learning for NLP, 2009.
  Cross-validation for accuracy is effective for semi-supervised learning on labeled data as small as 10 items. [pdf | slides]


• David Andrzejewski and Xiaojin Zhu. Latent Dirichlet allocation with topic-in-set knowledge. In NAACL 2009 Workshop on Semi-supervised Learning for NLP, 2009.
  [pdf | single thread code | parallel code for multi-core machines]


• Xiaojin Zhu, Zhiting Xu, and Tushar Khot. How creative is your writing? A linguistic creativity measure from computer science and cognitive psychology perspectives. In NAACL 2009 Workshop on Computational Approaches to Linguistic Creativity, 2009.
  Predict creativity of text using linear regression with features extracted from Google 1T 5gram corpus, WordNet, and Leuven word norms. [pdf] [data]


• Xiaojin Zhu, Andrew B. Goldberg, and Tushar Khot. Some new directions in graph-based semisupervised learning (invited paper). In IEEE International Conference on Multimedia and Expo (ICME), Special Session on Semi-Supervised Learning for Multimedia Analysis, 2009.
  A brief discussion on compressive sensing and graph-based semi-supervised learning. [pdf | slides]


• Andrew B. Goldberg, Jake Rosin, Xiaojin Zhu, and Charles R. Dyer. Toward Text-to-Picture Synthesis. In NIPS 2009 Symposium on Assistive Machine Learning for People with Disabilities, 2009.
  [pdf]

2008

• Rui Castro, Charles Kalish, Robert Nowak, Ruichen Qian, Timothy Rogers, and Xiaojin Zhu. Human active learning. In Advances in Neural Information Processing Systems (NIPS) 22, 2008.
  Can humans perform and benefit from active learning in categorization tasks? We conduct behavioral experiments and compare humans' learning rate to predictions by statistical learning theory. The short answer is Yes.
  [paper pdf | poster | 3D shapes]


• Aarti Singh, Robert Nowak, and Xiaojin Zhu. Unlabeled data: Now it helps, now it doesn't. In Advances in Neural Information Processing Systems (NIPS) 22, 2008. Oral presentation.
  Is semi-supervised learning (SSL) better than supervised learning (SL) in theory? We prove that as the distance between two classes gets closer and eventually overlap, there are several distinct phases in which SSL is better than SL, while in other phases SSL is no better than SL. [preprint: pdf | extended tech report | Errata | slides ]


• Andrew B. Goldberg, Ming Li, and Xiaojin Zhu. Online Manifold Regularization: A New Learning Setting and Empirical Study. In The European Conference on Machine Learning and Principles and Practice of Knowledge Discovery in Databases (ECML PKDD), 2008.
  Online semi-supervised learning. The key is stochastic gradient descent on any convex semi-supervised risk functional, with two practical approximations for manifold regularization: buffering and random projection tree. [pdf | poster | slides at Interface'08]


• Xiaojin Zhu, Andrew B. Goldberg, Michael Rabbat, and Robert Nowak. Learning bigrams from unigrams. In The 46th Annual Meeting of the Association for Computational Linguistics: Human Language Technologies (ACL), 2008.
  If I give you a text document in bag-of-word (unigram count vector) format, you will not know the order between words. What if I give you 10,000 documents, each in bag-of-word format? Surprisingly, we can partially recover a bigram language model just from these bag-of-word documents. [pdf | slides | discussion]


• Andrew B. Goldberg, Xiaojin Zhu, Charles R. Dyer, Mohamed Eldawy, and Lijie Heng. Easy as ABC? Facilitating pictorial communication via semantically enhanced layout. In Twelfth Conference on Computational Natural Language Learning (CoNLL), 2008.
  If you have pictures for individual words in a sentence, how do you compose them to best convey the meaning of the sentence? We learn an "ABC" layout using semantic role labeling and conditional random fields, and conduct a user study. [pdf | slides]


• Xiaojin Zhu, Michael Coen, Shelley Prudom, Ricki Colman, and Joseph Kemnitz. Online learning in monkeys. In Twenty-Third AAAI Conference on Artificial Intelligence (AAAI-08), 2008. (short paper)
  We compare rhesus monkeys playing the Wisconsin Card Sorting Task to online machine learning algorithms. [pdf | poster]


• Nathan Rosenblum, Xiaojin Zhu, Barton Miller, and Karen Hunt. Learning to analyze binary computer code. In Twenty-Third AAAI Conference on Artificial Intelligence (AAAI-08), 2008.
  An extended version of the NIPS07 workshop paper, including high throughput computing and a formal analysis of self-repairing disassembly. [pdf]


• Nathanael Fillmore, Andrew B. Goldberg, and Xiaojin Zhu. Document recovery from bag-of-word indices. Technical Report Computer Science TR1645, University of Wisconsin-Madison, 2008.
  Given a bag-of-words vector, recover the original ordered document. [pdf]

2007

• Nathan Rosenblum, Xiaojin Zhu, Barton Miller, and Karen Hunt. Machine Learning-Assisted Binary Code Analysis. In NIPS workshop on Machine Learning in Adversarial Environments for Computer Security, 2007.
  Identify function entry points in binary code using Markov Random Fields on both local instruction patterns and global control flow structures. [pdf]


• David Andrzejewski, Anne Mulhern, Ben Liblit, and Xiaojin Zhu. Statistical debugging using latent topic models. In Proceedings of the 18th European Conference on Machine Learning (ECML), 2007.
  Representing software execution traces using "bag-of-words", where the words are instrumented probes in the software. A Delta-Latent-Dirichlet-Allocation (ΔLDA) model to identify weak latent topics that correspond to distinct software bugs. [pdf | slides | code]


• Xiaojin Zhu, Timothy Rogers, Ruichen Qian, and Chuck Kalish. Humans perform semi-supervised classification too. In Twenty-Second AAAI Conference on Artificial Intelligence (AAAI-07), 2007.
  We show that humans determine class boundaries using both labeled and unlabeled data, just like certain semi-supervised machine learning models. [pdf | slides | poster | data]


• Xiaojin Zhu, Andrew Goldberg, Mohamed Eldawy, Charles Dyer, and Bradley Strock. A text-to-picture synthesis system for augmenting communication. In The Integrated Intelligence Track of the Twenty-Second AAAI Conference on Artificial Intelligence (AAAI-07), 2007.
  Synthesizing a picture from general, unrestricted natural language text, to convey the gist of the text. [pdf | slides]


• Xiaojin Zhu and Andrew Goldberg. Kernel regression with order preferences. In Twenty-Second AAAI Conference on Artificial Intelligence (AAAI-07), 2007.
  A linear program to incorporate order preferences ("I think the target value is larger at x1 than at x2") as regularizer in regression. [pdf | slides | TR 1578 version]


• Andrew Goldberg, Xiaojin Zhu, and Stephen Wright. Dissimilarity in graph-based semi-supervised classification. In Eleventh International Conference on Artificial Intelligence and Statistics (AISTATS), 2007.
  A convex quadratic program to incorporate cannot-links (two examples should have different labels) into binary and multiclass classification. Extends graph-based semi-supervised learning to mixed graphs. [pdf | poster]


• Xiaojin Zhu, Andrew Goldberg, Jurgen Van Gael, and David Andrzejewski. Improving diversity in ranking using absorbing random walks. In Human Language Technologies: The Annual Conference of the North American Chapter of the Association for Computational Linguistics (NAACL-HLT), 2007.
  A ranking algorithm (GRASSHOPPER) that is similar to PageRank but encourages diversity in top ranked items, by turning already ranked items into absorbing states to penalize remaining similar items. [pdf] [code]


• Mariyam Mirza, Joel Sommers, Paul Barford, and Xiaojin Zhu. A machine learning approach to TCP throughput prediction. In The International Conference on Measurement and Modeling of Computer Systems (ACM SIGMETRICS), 2007.
  Apply Support Vector Regression to predict Internet file transfer rate from measurable features of the network. [pdf]


• Jurgen Van Gael and Xiaojin Zhu. Correlation clustering for crosslingual link detection. In International Joint Conference on Artificial Intelligence (IJCAI), 2007.
  Cluster news articles in different languages by event. A practical implementation of correlation clustering that involves linear program chunking. [pdf][data]


• Gregory Druck, Chris Pal, Xiaojin Zhu, and Andrew McCallum. Semi-supervised classification with hybrid generative/discriminative methods. In The Thirteenth ACM SIGKDD International Conference on Knowledge Discovery and Data Mining (KDD), 2007.
  [pdf]


• Jordan Boyd-Graber, David Blei, and Xiaojin Zhu. A topic model for word sense disambiguation. In Conference on Empirical Methods in Natural Language Processing (EMNLP-CoNLL), 2007.
  [pdf]


• SaiSuresh Krishnakumaran and Xiaojin Zhu. Hunting elusive metaphors using lexical resources. In NAACL 2007 Workshop on Computational Approaches to Figurative Language, 2007.
  Identify "The soldier is a lion" as a metaphor by noting the lack of WordNet hyponym relationship between "soldier" and "lion". Extends to verb-noun or adjective-noun pairs using Google Web 1T bigram counts. [pdf] [data]

2006

• Xiaojin Zhu, Jaz Kandola, John Lafferty, and Zoubin Ghahramani. Graph kernels by spectral transforms. In O. Chapelle, B. Schölkopf, and A. Zien, editors, Semi-Supervised Learning. MIT Press, 2006.
  Keep the eigenvectors of a graph Laplacian, but optimize the eigenvalues under the constraints that smoother eigenvectors should have larger eigenvalues, to maximize kernel-target alignment on training data. Extended version of NIPS05 paper. [pdf]


• Andrew Goldberg and Xiaojin Zhu. Seeing stars when there aren't many stars: Graph-based semi-supervised learning for sentiment categorization. In HLT-NAACL 2006 Workshop on Textgraphs: Graph-based Algorithms for Natural Language Processing, New York, NY, 2006.
  Do people like a movie? We extend the classic Pang&Lee movie sentiment paper to semi-supervised learning by building a graph over labeled and unlabeled movie reviews. [pdf | slides]


• Andrew Goldberg, Dave Andrzejewski, Jurgen Van Gael, Burr Settles, Xiaojin Zhu, and Mark Craven. Ranking biomedical passages for relevance and diversity: University of Wisconsin, Madison at TREC Genomics 2006. In Proceedings of the Fifteenth Text Retrieval Conference (TREC), 2006.
  Our TREC06 biomedical passage retrieval system, focusing on query generation and result reranking with GRASSHOPPER (see the NAACL07 paper). [pdf | slides | poster]


• Xiaojin Zhu, David Blei, and John Lafferty. TagLDA: Bringing document structure knowledge into topic models. Technical Report 1553, Department of Computer Sciences, University of Wisconsin-Madison, 2006.
  An extension to Latent Dirichlet Allocation (LDA), where each topic can exhibits different word distributions in different parts (e.g., abstract, body, reference) of a document. Uses a factorized model to prevent combinatorial explosion. [pdf]

2005

• Xiaojin Zhu. Semi-supervised learning literature survey. Technical Report 1530, Department of Computer Sciences, University of Wisconsin, Madison, 2005.
  We review the literature on semi-supervised learning, i.e., machine learning from both labeled and unlabeled data. This online paper is updated frequently to incorporate the latest development in the field. [pdf]


• Xiaojin Zhu. Semi-Supervised Learning with Graphs. PhD thesis, Carnegie Mellon University, 2005. CMU-LTI-05-192. [pdf]
  My Ph.D. thesis on graph-based semi-supervised learning, including label propagation, Gaussian random fields and harmonic functions, semi-supervised active learning, graph hyperparameter learning, kernel matrices from graph Laplacian, sparse representation and so on.


• Xiaojin Zhu and John Lafferty. Harmonic mixtures: combining mixture models and graph-based methods for inductive and scalable semi-supervised learning. In The 22nd International Conference on Machine Learning (ICML). ACM Press, 2005.
  Making graph-based semi-supervised learning faster and handling unseen data, by first modeling data with a mixture model (e.g., GMM), then treating mixture components (instead of individual data points) as nodes in the graph. [pdf][small teapot data (.mat)]


• Xiaojin Zhu, Zoubin Ghahramani, and John Lafferty. Time-sensitive Dirichlet process mixture models. Technical Report CMU-CALD-05-104, Carnegie Mellon University, 2005.
  An extension to Dirichlet Process Mixture Models where the probability of a state depends not only on the numbers of previous states, but also on how old they are. [pdf]


• Maria-Florina Balcan, Avrim Blum, Patrick Pakyan Choi, John Lafferty, Brian Pantano, Mugizi Robert Rwebangira, and Xiaojin Zhu. Person identification in webcam images: An application of semi-supervised learning. In ICML 2005 Workshop on Learning with Partially Classified Training Data, 2005.
  Use abundant unlabeled frames to improve people recognition by Webcam. The graph over Webcam image frames uses close-in-time edges, foreground color histogram edges (people with similar apparel), and similar-face edges. [pdf] [FreeFoodCam dataset (.tgz 335MB)]


• Xiaojin Zhu, Jaz Kandola, Zoubin Ghahramani, and John Lafferty. Nonparametric transforms of graph kernels for semi-supervised learning. In Lawrence K. Saul, Yair Weiss, and Léon Bottou, editors, Advances in Neural Information Processing Systems (NIPS) 17. MIT Press, Cambridge, MA, 2005.
  Keep the eigenvectors of a graph Laplacian, but optimize the eigenvalues under the constraints that smoother eigenvectors should have larger eigenvalues, to maximize kernel-target alignment on training data. [pdf] [Matlab code & data] [QP notes]

2004

• John Lafferty, Xiaojin Zhu, and Yan Liu. Kernel conditional random fields: Representation and clique selection. In The 21st International Conference on Machine Learning (ICML), 2004.
  We kernelize Conditional Random Fields, which is an alternative to Maximum Margin Markov Networks. We propose greedy clique selection in the dual for sparse representation. [ps] [pdf]

2003

• Xiaojin Zhu, Zoubin Ghahramani, and John Lafferty. Semi-supervised learning using Gaussian fields and harmonic functions. In The 20th International Conference on Machine Learning (ICML), 2003. ICML 10-Year Classic Paper Prize.
  A graph-based semi-supervised learning algorithm that creates a graph over labeled and unlabeled examples. More similar examples are connected by edges with higher weights. The intuition is for the labels to propagate on the graph to unlabeled data. The solution can be found with simple matrix operations, and has strong connections to spectral graph theory. [ps.gz] [pdf] [Matlab code] [data]


• Xiaojin Zhu, John Lafferty, and Zoubin Ghahramani. Combining active learning and semi-supervised learning using Gaussian fields and harmonic functions. In ICML 2003 workshop on The Continuum from Labeled to Unlabeled Data in Machine Learning and Data Mining, 2003.
  Actively selects an unlabeled point to ask for the label, by minimizing an estimated classification error (instead of simply picking the most ambiguous unlabeled point). Once the label is obtained, efficiently retrain the classifier with both labeled and unlabeled data. [ps.gz] [pdf] [Matlab code]


• Xiaojin Zhu, John Lafferty, and Zoubin Ghahramani. Semi-supervised learning: From Gaussian fields to Gaussian processes. Technical Report CMU-CS-03-175, Carnegie Mellon University, 2003.
  We establish the connection between the inverse graph Laplacian and kernel Gram matrix, and learn hyperparameters for graph weights with evidence maximization. However, this is not a true Gaussian process since unseen points (not in training labeled and unlabeled data) are not handled well. [ps.gz] [pdf]

2002

• Xiaojin Zhu and Zoubin Ghahramani. Learning from labeled and unlabeled data with label propagation. Technical Report CMU-CALD-02-107, Carnegie Mellon University, 2002.
  Precursor of the ICML03 paper. The intuition of label propagation is introduced, together with an iterative algorithm which amounts to relaxation method. [ps.gz] [pdf]


• Xiaojin Zhu and Zoubin Ghahramani. Towards semi-supervised classification with Markov random fields. Technical Report CMU-CALD-02-106, Carnegie Mellon University, 2002.
  Yet another precursor of the ICML03 paper. The graph is defined, but as a Boltzmann machines (discrete states) rather than the later Gaussian random fields (continuous states). Inference, with MCMC, is difficult. [ps.gz] [pdf]

2001

• Ronald Rosenfeld, Stanley Chen, and Xiaojin Zhu. Whole-sentence exponential language models: a vehicle for linguistic-statistical integration. Computers Speech and Language, 15(1), 2001.
  Directly model the probability of a sentence with an exponential model, instead of using the chain rule on words. Can use arbitrary, long range features. [pdf]


• Stefanie Shriver, Arthur Toth, Xiaojin Zhu, Alex Rudnicky, and Roni Rosenfeld. A unified design for human-machine voice interaction. In Human Factors in Computing Systems (CHI). ACM Press, 2001.
  In order for humans to use speech interfaces, they might need to learn how to speak to machines. [ps]


• Xiaojin Zhu and Ronald Rosenfeld. Improving trigram language modeling with the World Wide Web. In Proceedings of the International Conference on Acoustics, Speech and Signal Processing (ICASSP), 2001.
  Estimating n-gram probabilities by submitting word sequences as phrase queries to search engines. [pdf] [tech report version CMU-CS-00-171 ps]

2000

• Ronald Rosenfeld, Xiaojin Zhu, Stefanie Shriver, Arthur Toth, Kevin Lenzo, and Alan Black. Towards a universal speech interface. In International Conference on Spoken Language Processing (ICSLP), 2000.
  A general speech input paradigm that attempts to structurize human speech to facilitate speech recognition. [pdf]


• Xiaojin Zhu, Jie Yang, and Alex Waibel. Segmenting hands of arbitrary color. In Fourth IEEE International Conference on Automatic Face and Gesture Recognition, 2000.
  We model the color histogram of a scene by a Gaussian mixture model, one of the mixture component is the hand. [ps.gz]

1999

• Jie Yang, Xiaojin Zhu, Ralph Gross, John Kominek, Yue Pan, and Alex Waibel. Multimodal people ID for multimedia meeting browser. In The Seventh ACM International Multimedia Conference, 1999.
  Use face recognition, speaker identification, color histogram, and sound direction to identify meeting participants. [link]


• Xiaojin Zhu, Stanley F. Chen, and Ronald Rosenfeld. Linguistic features for whole sentence maximum entropy language models. In Proceedings of the 5th European Conference on Speech Communication and Technology (Eurospeech), 1999.
  Parse a real corpus and a trigram-generated corpus using a shallow parser. Identify features that behave differently in the two corpora. Use them to build a better language model. [ps]