Description |
viii, 348 pages : illustrations ; 26 cm |
Series |
Neural information processing series |
|
Advances in neural information processing systems |
|
Neural information processing series.
|
Contents |
1 Measuring Similarity with Kernels -- 1.1 Introduction -- 1.2 Kernels -- 1.3 Operating in Reproducing Kernel Hilbert Spaces -- 1.4 Kernels for Structured Data -- 1.5 An Example of a Structured Prediction Algorithm Using Kernels -- 2 Discriminative Models -- 2.1 Introduction -- 2.2 Online Large-Margin Algorithms -- 2.3 Support Vector Estimation -- 2.4 Margin-Based Loss Functions -- 2.5 Margins and Uniform Convergence Bounds -- 3 Modeling Structure via Graphical Models -- 3.1 Introduction -- 3.2 Conditional Independence -- 3.3 Markov Networks -- 3.4 Bayesian Networks -- 3.5 Inference Algorithms -- 3.6 Exponential Families -- 3.7 Probabilistic Context-Free Grammars -- 3.8 Structured Prediction -- Structured Prediction Based on Discriminative Models -- 4 Joint Kernel Maps -- 4.1 Introduction -- 4.2 Incorporating Correlations into Linear Regression -- 4.3 Linear Maps and Kernel Methods: Generalizing Support Vector Machines -- 4.4 Joint Kernel Maps -- 4.5 Joint Kernels -- 4.6 Experiments -- 5 Support Vector Machine Learning for Interdependent and Structured Output Spaces -- 5.1 Introduction -- 5.2 A Framework for Structured/Interdependent Output Learning -- 5.3 A Maximum-Margin Formulation -- 5.4 Cutting-Plane Algorithm -- 5.5 Alternative Margin Formulations -- 5.6 Experiments -- 6 Efficient Algorithms for Max-Margin Structured Classification -- 6.1 Introduction -- 6.2 Structured Classification Model -- 6.3 Efficient Optimization on the Marginal Dual Polytope -- 6.4 Experiments -- 7 Discriminative Learning of Prediction Suffix Trees with the Perceptron Algorithm -- 7.1 Introduction -- 7.2 Suffix Trees for Stream Prediction -- 7.3 PSTs as Separating Hyperplanes and the perceptron Algorithm -- 7.4 The Self-Bounded Perceptron for PST Learning -- 8 A General Regression Framework for Learning String-to-String Mappings -- 8.1 Introduction -- 8.2 General Formulation -- 8.3 Regression Problems and Algorithms -- 8.4 Pre-Image Solution for Strings -- 8.5 Speeding up Training -- 8.6 Comparison with Other Algorithms -- 8.7 Experiments -- 9 Learning as Search Optimization -- 9.1 Introduction -- 9.2 Previous Work -- 9.3 Search Optimization -- 9.4 Experiments -- 10 Energy-Based Models -- 10.1 Introduction -- 10.2 Energy-Based Training: Architecture and Loss Function -- 10.3 Simple Architectures -- 10.4 Latent Variable Architectures -- 10.5 Analysis of Loss Functions for Energy-Based Models -- 10.6 Efficient Inference: Nonprobabilistic Factor Graphs -- 10.7 EBMs for Sequence Labeling and Structured Outputs -- 11 Generalization Bounds and Consistency for Structured Labeling -- 11.1 Introduction -- 11.2 PAC-Bayesian Generalization Bounds -- 11.3 Hinge Loss -- 11.4 Consistency -- 11.5 A Generalization of Theorem 62 -- 11.6 Proofs of Theorems 61 and 62 -- Structured Prediction Using Probabilistic Models -- 12 Kernel Conditional Graphical Models -- 12.1 Introduction -- 12.2 A Unifying Review -- 12.3 Conditional Graphical Models -- 12.4 Experiments -- 13 Density Estimation of Structured Outputs in Reproducing Kernel Hilbert Spaces -- 13.1 Introduction -- 13.2 Estimating Conditional Probability Distributions over Structured Outputs -- 13.3 A Sparse Greedy Optimization -- 13.4 Experiments: Sequence Labeling -- 14 Gaussian Process Belief Propagation -- 14.1 Introduction -- 14.2 Data and Model Dimension -- 14.3 Semiparametric Latent Factor Models -- 14.4 Gaussian Process Belief Propagation -- 14.5 Parameter Learning |
Notes |
Collected papers based on talks presented at two Neural Information Processing Systems workshops |
Bibliography |
Includes bibliographical references (pages [319]-340) and index |
Subject |
Machine learning.
|
|
Computer algorithms.
|
|
Kernel functions.
|
|
Data structures (Computer science)
|
Author |
BakIr, Gökhan.
|
|
Neural Information Processing Systems Foundation.
|
LC no. |
2006047001 |
ISBN |
9780262026178 alkaline paper |
|
0262026171 alkaline paper |
|