Proceedings of the Twentieth International Conference on Machine Learning
Edited by Tom Fawcett and Nina Mishra
Washington, DC, August 21-24, 2003. Published by The AAAI Press, Menlo Park, California. This proceedings is also available in book format.
Please Note: Abstracts are linked to individual titles, and will appear in a separate browser window. Full-text versions of the papers are linked to the abstract text. Access to full text may be restricted to AAAI members. PDF file sizes may be large!
Contents
Preface / xiii
Tom Fawcett and Nina Mishra
ICML 2003 Corporate Sponsors / xviii
Hidden Markov Support Vector Machines / 3
Yasemin Altun, Ioannis Tsochantaridis, and Thomas Hofmann
Learning Distance Functions using Equivalence Relations / 11
Aharon Bar-Hillel, Tomer Hertz, Noam Shental, and Daphna Weinshall
Online Choice of Active Learning Algorithms / 19
Yoram Baram, Ran El-Yaniv, and Kobi Luz
Learning Logic Programs for Layout Analysis Correction / 27
Margherita Berardi, Michelangelo Ceci, Floriana Esposito, and Donato Malerba
Multi-Objective Programming in SVMs / 35
Jinbo Bi
Regression Error Characteristic Curves / 43
Jinbo Bi and Kristin P. Bennett
Choosing Between Two Learning Algorithms Based on Calibrated Tests / 51
Remco R. Bouckaert
Incorporating Diversity in Active Learning with Support Vector Machines / 59
Klaus Brinker
The Use of the Ambiguity Decomposition in Neural Network Ensemble Learning Methods / 67
Gavin Brown and Jeremy Wyatt
Tractable Bayesian Learning of Tree Augmented Naive Bayes Models / 75
Jesüs Cerquides and Ramon López de Màntaras
AWESOME: A General Multiagent Learning Algorithm that Converges in Self-Play and Learns a Best Response Against Stationary Opponents / 83
Vincent Conitzer and Tuomas Sandholm
BL-WoLF: A Framework For Loss-Bounded Learnability In Zero-Sum Games / 91
Vincent Conitzer and Tuomas Sandholm
Semi-Supervised Learning of Mixture Models / 99
Fabio Gagliardi Cozman, Ira Cohen, and Marcelo Cesar Cirelo
On Kernel Methods for Relational Learning / 107
Chad Cumby and Dan Roth
Fast Query-Optimized Kernel Machine Classification Via Incremental Approximate Nearest Support Vectors / 115
Dennis DeCoste and Dominic Mazzoni
Relational Instance Based Regression for Relational Reinforcement Learning / 123
Kurt Driessens and Jan Ramon
Design for an Optimal Probe / 131
Michael Duff
Diffusion Approximation for Bayesian Markov Chains / 139
Michael Duff
Using the Triangle Inequality to Accelerate k-Means / 147
Charles Elkan
Bayes Meets Bellman: The Gaussian Process Approach to Temporal Difference Learning / 154
Yaakov Engel, Shie Mannor, and Ron Meir
Action Elimination and Stopping Conditions for Reinforcement Learning / 162
Eyal Even-Dar, Shie Mannor, and Yishay Mansour
Utilizing Domain Knowledge in Neuroevolution / 170
James Fan, Raymond Lau, and Risto Miikkulainen
Boosting Lazy Decision Trees / 178
Xiaoli Zhang Fern and Carla E. Brodley
Random Projection for High Dimensional Data Clustering: A Cluster Ensemble Approach / 186
Xiaoli Zhang Fern and Carla E. Brodley
The Geometry of ROC Space: Understanding Machine Learning Metrics through ROC Isometrics / 194
Peter A. Flach
An Analysis of Rule Evaluation Metrics / 202
Johannes Fürnkranz and Peter A. Flach
Margin Distribution and Learning / 210
Ashutosh Garg and Dan Roth
Perceptron Based Learning with Example Dependent and Noisy Costs / 218
Peter Geibel and Fritz Wysotzki
Hierarchical Policy Gradient Algorithms / 226
Mohammad Ghavamzadeh and Sridhar Mahadevan
Solving Noisy Linear Operator Equations by Gaussian Processes: Application to Ordinary and Partial Differential Equations / 234
Thore Graepel
Correlated Q-Learning / 242
Amy Greenwald and Keith Hall
Online Ranking/Collaborative Filtering Using the Perceptron Algorithm / 250
Edward F. Harrington
Goal-directed Learning to Fly / 258
Andrew Isaac and Claude Sammut
Probabilistic Classifiers and the Concepts They Recognize / 266
Manfred Jaeger
Avoiding Bias when Aggregating Relational Data with Degree Disparity / 274
David Jensen, Jennifer Neville, and Michael Hay
A Faster Iterative Scaling Algorithm for Conditional Exponential Model / 282
Rong Jin, Rong Yan, Jian Zhang, and Alex G. Hauptmann
Transductive Learning via Spectral Graph Partitioning / 290
Thorsten Joachims
Evolving Strategies for Focused Web Crawling / 298
Judy Johnson, Kostas Tsioutsiouliklis, and C. Lee Giles
Exploration in Metric State Spaces / 306
Sham Kakade, Michael Kearns, and John Langford
Representational Issues in Meta-Learning / 313
Alexandros Kalousis and Melanie Hilario
Marginalized Kernels Between Labeled Graphs / 321
Hisashi Kashima, Koji Tsuda, and Akihiro Inokuchi
Informative Discriminant Analysis / 329
Samuel Kaski and Jaakko Peltonen
Characteristics of Long-term Learning in Soar and its Application to the Utility Problem / 337
William G. Kennedy and Kenneth A. De Jong
Unsupervised Learning with Permuted Data / 345
Sergey Kirshner, Sridevi Parise, and Padhraic Smyth
Discriminative Gaussian Mixture Models: A Comparison with Kernel Classifiers / 353
Aldebaro Klautau, Nikola Jevtic, and Alon Orlitsky
A Kernel Between Sets of Vectors / 361
Risi Kondor and Tony Jebara
The Significance of Temporal-Difference Learning in Self-Play Training td-Rummy versus EVO-rummy / 369
Clifford Kotnik and Jugal Kalita
Visual Learning by Evolutionary Feature Synthesis / 376
Krzysztof Krawiec and Bir Bhanu
Classification of Text Documents Based on Minimum System Entropy / 384
Raghu Krishnapuram, Krishna P. Chitrapura, and Sachindra Joshi
Finding Underlying Connections: A Fast Graph-Based Method for Link Analysis and Collaboration Queries / 392
Jeremy Kubica, Andrew Moore, David Cohn, and Jeff Schneider
Learning with Idealized Kernels / 400
James T. Kwok and Ivor W. Tsang
The Pre-Image Problem in Kernel Methods / 408
James Kwok and Ivor Tsang
Improving Accuracy and Cost of Two-class and Multi-class Probabilistic Classifiers Using ROC Curves / 416
Nicolas Lachiche and Peter Flach
Reinforcement Learning as Classification: Leveraging Modern Classifiers / 424
Michail Lagoudakis and Ronald Parr
Robust Induction of Process Models from Time-Series Data / 432
Pat Langley, Dileep George, Stephen Bay, and Kazumi Saito
The Influence of Reward on the Speed of Reinforcement Learning: An Analysis of Shaping / 440
Adam Laud and Gerald DeJong
Learning with Positive and Unlabeled Examples Using Weighted Logistic Regression / 448
Wee Sun Lee and Bing Liu
Linear Programming Boosting for Uneven Datasets / 456
Jurij Leskovec and John Shawe-Taylor
Text Classification Using Stochastic Keyword Generation / 464
Cong Li, Ji-Rong Wen, and Hang Li
A Loss Function Analysis for Classification Methods in Text Categorization / 472
Fan Li and Yiming Yang
Decision Tree with Better Ranking / 480
Charles X. Ling and Robert J. Yan
An Evaluation on Feature Selection for Text Clustering / 488
Tao Liu, Shengping Liu, Zheng Chen, and Wei-Ying Ma
Link-based Classification / 496
Qing Lu and Lise Getoor
Hierarchical Latent Knowledge Analysis for Co-occurrence Data / 504
Hiroshi Mamitsuka
The Cross Entropy Method for Fast Policy Search / 512
Shie Mannor, Reuven Rubinstein, and Yohai Gat
The Set Covering Machine with Data-Dependent Half-Spaces / 520
Mario Marchand, Mohak Shah, John Shawe-Taylor, and Marina Sokolova
Identifying Predictive Structures in Relational Data Using Multiple Instance Learning / 528
Amy McGovern and David Jensen
Planning in the Presence of Cost Functions Controlled by an Adversary / 536
H. Brendan McMahan, Geoffrey J. Gordon, and Avrim Blum
Using Linear-threshold Algorithms to Combine Multi-class Sub-experts / 544
Chris Mesterharm
Optimal Reinsertion: A New Search Operator for Accelerated and More Accurate Bayesian Network Structure Learning / 552
Andrew Moore and Weng-Keen Wong
Error Bounds for Approximate Policy Iteration / 560
Rémi Munos
Machine Learning with Hyperkernels / 568
Cheng Soon Ong and Alexander J. Smola
Justification-based Multiagent Learning / 576
Santi Ontañón and Enric Plaza
Mixtures of Conditional Maximum Entropy Models / 584
Dmitry Pavlov, Alexandrin Popescul, David M. Pennock, and Lyle H. Ungar
Online Feature Selection using Grafting / 592
Simon Perkins and James Theiler
Weighted Order Statistic Classifiers with Large Rank-Order Margin / 600
Reid Porter, Damian Eads, Don Hush, and James Theiler
Relativized Options: Choosing the Right Transformation / 608
Balaraman Ravindran and Andrew G. Barto
Tackling the Poor Assumptions of Naive Bayes Text Classifiers / 616
Jason D. Rennie, Lawrence Shih, Jaime Teevan, and David Karger
Learning with Knowledge from Multiple Experts / 624
Matthew Richardson and Pedro Domingos
Combining td-learning with Cascade-correlation Networks / 632
François Rivest and Doina Precup
Kernel PLS-SVC for Linear and Nonlinear Classification / 640
Roman Rosipal, Leonard J. Trejo, and Bryan Matthews
Stochastic Local Search in k-Term DNF Learning / 648
Ulrich Rückert and Stefan Kramer
Q-Decomposition for Reinforcement Learning Agents / 656
Stuart Russell and Andrew L. Zimdars
Adaptive Overrelaxed Bound Optimization Methods / 664
Ruslan Salakhutdinov and Sam Roweis
Optimization with EM and Expectation-Conjugate-Gradient / 672
Ruslan Salakhutdinov, Sam Roweis, and Zoubin Ghahramani
td(0) Converges Provably Faster than the Residual Gradient Algorithm / 680
Ralf Schoknecht and Artur Merke
On State Merging in Grammatical Inference: A Statistical Approach for Dealing with Noisy Data / 688
Marc Sebban and Jean-Christophe Janodet
Text Bundling: Statistics Based Data-Reduction / 696
Lawrence Shih, Jason Rennie, Yu-Han Chang, and David Karger
Flexible Mixture Model for Collaborative Filtering / 704
Luo Si and Rong Jin
Learning Predictive State Representations / 712
Satinder Singh, Michael L. Littman, Nicholas K. Jong, David Pardoe, and Peter Stone
Weighted Low-Rank Approximations / 720
Nathan Srebro and Tommi Jaakkola
Learning To Cooperate in a Social Dilemma: A Satisficing Approach to Bargaining / 728
Jeffrey L. Stimpson and Michael A. Goodrich
Evolutionary MCMC Sampling and Optimization in Discrete Spaces / 736
Malcolm J. A. Strens
Learning on the Test Data: Leveraging Unseen Features / 744
Ben Taskar, Ming Fai Wong, and Daphne Koller
Low Bias Bagged Support Vector Machines / 752
Giorgio Valentini and Thomas G. Dietterich
SimpleSVM / 760
S. V. N. Vishwanathan, Alexander J. Smola, and M. Narashima Murty
Testing Exchangeability On-Line / 768
Vladimir Vovk, Ilia Nouretdinov, and Alex Gammerman
Model-based Policy Gradient Reinforcement Learning / 776
Xin Wang and Thomas G. Dietterich
Learning Mixture Models with the Latent Maximum Entropy Principle / 784
Shaojun Wang, Dale Schuurmans, Fuchun Peng, and Yunxin Zhao
Principled Methods for Advising Reinforcement Learning Agents / 792
Eric Wiewiora, Garrison Cottrell, and Charles Elkan
DISTILL: Learning Domain-Specific Planners by Example / 800
Elly Winner and Manuela Veloso
Bayesian Network Anomaly Pattern Detection for Disease Outbreaks / 808
Weng-Keen Wong, Andrew Moore, Gregory Cooper, and Michael Wagner
Adaptive Feature-Space Conformal Transformation for Imbalanced-Data Learning / 816
Gang Wu and Edward Y. Chang
New v-Support Vector Machines and their Sequential Minimal Optimization / 824
Xiaoyun Wu and Rohini Srihari
Cross-Entropy Directed Embedding of Network Data / 832
Takeshi Yamada, Kazumi Saito, and Naonori Ueda
Decision-tree Induction from Time-series Data Based on a Standard-example Split Test / 840
Yuu Yamada, Einoshin Suzuki, Hideto Yokoi, and Katsuhiko Takabayashi
Optimizing Classifier Performance via an Approximation to the Wilcoxon-Mann-Whitney Statistic / 848
Lian Yan, Robert Dodier, Michael C. Mozer, and Richard Wolniewicz
Feature Selection for High-Dimensional Data: A Fast Correlation-Based Filter Solution / 856
Lei Yu and Huan Liu
Isometric Embedding and Continuum ISOMAP / 864
Hongyuan Zha, and Zhenyue Zhang
Learning Metrics via Discriminant Kernels and Multidimensional Scaling: Toward Expected Euclidean Representation / 872
Zhihua Zhang
Learning from Attribute Value Taxonomies and Partially Specified Instances / 880
Jun Zhang and Vasant Honavar
Modified Logistic Regression: An Approximation to SVM and Its Applications in Large-Scale Text Categorization / 888
Jian Zhang, Rong Jin, Yiming Yang, and Alex G. Hauptmann
Exploration and Exploitation in Adaptive Filtering Based on Bayesian Active Learning / 896
Yi Zhang, Wei Xu, and Jamie Callan
On the Convergence of Boosting Procedures / 904
Tong Zhang and Bin Yu
Semi-Supervised Learning Using Gaussian Fields and Harmonic Functions / 912
Xiaojin Zhu, Zoubin Ghahramani, and John Lafferty
Eliminating Class Noise in Large Datasets / 920
Xingquan Zhu, Xindong Wu, and Qijun Chen
Online Convex Programming and Generalized Infinitesimal Gradient Ascent / 928
Martin Zinkevich
Index / 937