Preface......Page 13
The Data......Page 15
Unsupervised Learning......Page 16
Supervised Learning......Page 17
Learning from Both Labeled and Unlabeled Data......Page 23
How is Semi-Supervised Learning Possible?......Page 25
Inductive vs. Transductive Semi-Supervised Learning......Page 26
Caveats......Page 27
Self-Training Models......Page 29
Mixture Models for Supervised Classification......Page 35
Mixture Models for Semi-Supervised Classification......Page 39
Optimization with the EM Algorithm......Page 40
The Assumptions of Mixture Models......Page 42
Other Issues in Generative Models......Page 44
Cluster-then-Label Methods......Page 45
Two Views of an Instance......Page 49
Co-Training......Page 50
The Assumptions of Co-Training......Page 51
Multiview Learning......Page 52
The Graph......Page 57
Mincut......Page 59
Harmonic Function......Page 60
Manifold Regularization......Page 64
The Assumption of Graph-Based Methods......Page 65
Support Vector Machines......Page 71
Semi-Supervised Support Vector Machines......Page 75
Entropy Regularization......Page 77
The Assumption of S3VMs and Entropy Regularization......Page 79
From Machine Learning to Cognitive Science......Page 83
Study One: Humans Learn from Unlabeled Test Data......Page 84
Study Two: Presence of Human Semi-Supervised Learning in a Simple Task......Page 86
Study Three: Absence of Human Semi-Supervised Learning in a Complex Task......Page 89
Discussions......Page 91
A Simple PAC Bound for Supervised Learning......Page 93
A Simple PAC Bound for Semi-Supervised Learning......Page 95
Future Directions of Semi-Supervised Learning......Page 97
Basic Mathematical Reference......Page 99
Semi-Supervised Learning Software......Page 103
Symbols......Page 107
Biography......Page 127
Index......Page 129