| http://dbpedia.org/ontology/abstract
|
In machine learning, the kernel embedding … In machine learning, the kernel embedding of distributions (also called the kernel mean or mean map) comprises a class of nonparametric methods in which a probability distribution is represented as an element of a reproducing kernel Hilbert space (RKHS). A generalization of the individual data-point feature mapping done in classical kernel methods, the embedding of distributions into infinite-dimensional feature spaces can preserve all of the statistical features of arbitrary distributions, while allowing one to compare and manipulate distributions using Hilbert space operations such as inner products, distances, projections, linear transformations, and spectral analysis. This learning framework is very general and can be applied to distributions over any space on which a sensible kernel function (measuring similarity between elements of ) may be defined. For example, various kernels have been proposed for learning from data which are: vectors in , discrete classes/categories, strings, graphs/networks, images, time series, manifolds, dynamical systems, and other structured objects. The theory behind kernel embeddings of distributions has been primarily developed by Alex Smola, Le Song , Arthur Gretton, and Bernhard Schölkopf. A review of recent works on kernel embedding of distributions can be found in. The analysis of distributions is fundamental in machine learning and statistics, and many algorithms in these fields rely on information theoretic approaches such as entropy, mutual information, or Kullback–Leibler divergence. However, to estimate these quantities, one must first either perform density estimation, or employ sophisticated space-partitioning/bias-correction strategies which are typically infeasible for high-dimensional data. Commonly, methods for modeling complex distributions rely on parametric assumptions that may be unfounded or computationally challenging (e.g. Gaussian mixture models), while nonparametric methods like kernel density estimation (Note: the smoothing kernels in this context have a different interpretation than the kernels discussed here) or characteristic function representation (via the Fourier transform of the distribution) break down in high-dimensional settings. Methods based on the kernel embedding of distributions sidestep these problems and also possess the following advantages: 1.
* Data may be modeled without restrictive assumptions about the form of the distributions and relationships between variables 2.
* Intermediate density estimation is not needed 3.
* Practitioners may specify the properties of a distribution most relevant for their problem (incorporating prior knowledge via choice of the kernel) 4.
* If a characteristic kernel is used, then the embedding can uniquely preserve all information about a distribution, while thanks to the kernel trick, computations on the potentially infinite-dimensional RKHS can be implemented in practice as simple Gram matrix operations 5.
* Dimensionality-independent rates of convergence for the empirical kernel mean (estimated using samples from the distribution) to the kernel embedding of the true underlying distribution can be proven. 6.
* Learning algorithms based on this framework exhibit good generalization ability and finite sample convergence, while often being simpler and more effective than information theoretic methods Thus, learning via the kernel embedding of distributions offers a principled drop-in replacement for information theoretic approaches and is a framework which not only subsumes many popular methods in machine learning and statistics as special cases, but also can lead to entirely new learning algorithms. lead to entirely new learning algorithms.
|
| http://dbpedia.org/ontology/wikiPageExternalLink
|
http://www.gatsby.ucl.ac.uk/~gretton/ +
, http://alex.smola.org/ +
, https://bitbucket.org/szzoli/ite/ +
, http://www.cc.gatech.edu/~lsong/ +
|
| http://dbpedia.org/ontology/wikiPageID
|
41370976
|
| http://dbpedia.org/ontology/wikiPageLength
|
53783
|
| http://dbpedia.org/ontology/wikiPageRevisionID
|
1073073352
|
| http://dbpedia.org/ontology/wikiPageWikiLink
|
http://dbpedia.org/resource/Category:Theory_of_probability_distributions +
, http://dbpedia.org/resource/Minimax_estimator +
, http://dbpedia.org/resource/Well-posed_problem +
, http://dbpedia.org/resource/Kernel_methods +
, http://dbpedia.org/resource/Vector_%28mathematics_and_physics%29 +
, http://dbpedia.org/resource/Gramian_matrix +
, http://dbpedia.org/resource/Bregman_divergence +
, http://dbpedia.org/resource/Standard_basis +
, http://dbpedia.org/resource/Bounded_function +
, http://dbpedia.org/resource/Network_theory +
, http://dbpedia.org/resource/Characteristic_function_%28probability_theory%29 +
, http://dbpedia.org/resource/Basis_function +
, http://dbpedia.org/resource/Entropy_%28information_theory%29 +
, http://dbpedia.org/resource/Conditional_distribution +
, http://dbpedia.org/resource/Sinc +
, http://dbpedia.org/resource/Markov_Random_Field +
, http://dbpedia.org/resource/Bernhard_Sch%C3%B6lkopf +
, http://dbpedia.org/resource/Entropy_estimation +
, http://dbpedia.org/resource/Domain_adaptation +
, http://dbpedia.org/resource/Kernel_density_estimation +
, http://dbpedia.org/resource/Continuous_function +
, http://dbpedia.org/resource/Tensor_product +
, http://dbpedia.org/resource/Projection_%28linear_algebra%29 +
, http://dbpedia.org/resource/Tikhonov_regularization +
, http://dbpedia.org/resource/Separable_space +
, http://dbpedia.org/resource/Support_vector_machine +
, http://dbpedia.org/resource/Hyperparameter +
, http://dbpedia.org/resource/Independence_%28probability_theory%29 +
, http://dbpedia.org/resource/Location_parameter +
, http://dbpedia.org/resource/Hilbert_space +
, http://dbpedia.org/resource/Gaussian_process +
, http://dbpedia.org/resource/Mixture_model +
, http://dbpedia.org/resource/Belief_propagation +
, http://dbpedia.org/resource/String_%28computer_science%29 +
, http://dbpedia.org/resource/Fourier_transform +
, http://dbpedia.org/resource/Incomplete_Cholesky_factorization +
, http://dbpedia.org/resource/Spectral_theory +
, http://dbpedia.org/resource/Concentration_of_measure +
, http://dbpedia.org/resource/Joint_probability_distribution +
, http://dbpedia.org/resource/Markov_random_field +
, http://dbpedia.org/resource/Support_%28mathematics%29 +
, http://dbpedia.org/resource/Orthogonal_matrix +
, http://dbpedia.org/resource/Tensor +
, http://dbpedia.org/resource/Kernel_trick +
, http://dbpedia.org/resource/Kronecker_delta +
, http://dbpedia.org/resource/Point_estimation +
, http://dbpedia.org/resource/Dimensionality_reduction +
, http://dbpedia.org/resource/Independent_and_identically_distributed_random_variables +
, http://dbpedia.org/resource/Radial_basis_function +
, http://dbpedia.org/resource/Compact_space +
, http://dbpedia.org/resource/Linear_subspace +
, http://dbpedia.org/resource/Category:Machine_learning +
, http://dbpedia.org/resource/Probability_distribution +
, http://dbpedia.org/resource/Pearson_correlation +
, http://dbpedia.org/resource/Hidden_Markov_Model +
, http://dbpedia.org/resource/Multiple-instance_learning +
, http://dbpedia.org/resource/Graphical_model +
, http://dbpedia.org/resource/Scale_parameter +
, http://dbpedia.org/resource/Entropy +
, http://dbpedia.org/resource/Dense_set +
, http://dbpedia.org/resource/Hilbert%E2%80%93Schmidt_integral_operator +
, http://dbpedia.org/resource/Overfitting +
, http://dbpedia.org/resource/Hilbert%E2%80%93Schmidt_operator +
, http://dbpedia.org/resource/Statistics +
, http://dbpedia.org/resource/Identity_matrix +
, http://dbpedia.org/resource/Graph_%28discrete_mathematics%29 +
, http://dbpedia.org/resource/Bias_of_an_estimator +
, http://dbpedia.org/resource/Cross-covariance +
, http://dbpedia.org/resource/Cluster_analysis +
, http://dbpedia.org/resource/Machine_learning +
, http://dbpedia.org/resource/Conditional_random_fields +
, http://dbpedia.org/resource/Hidden_Markov_model +
, http://dbpedia.org/resource/Manifold +
, http://dbpedia.org/resource/Kernel_principal_component_analysis +
, http://dbpedia.org/resource/Regularization_%28mathematics%29 +
, http://dbpedia.org/resource/Similarity_measure +
, http://dbpedia.org/resource/Curse_of_dimensionality +
, http://dbpedia.org/resource/Dynamical_systems +
, http://dbpedia.org/resource/Mutual_information +
, http://dbpedia.org/resource/Kernel_function +
, http://dbpedia.org/resource/Nonparametric +
, http://dbpedia.org/resource/Exponential_family +
, http://dbpedia.org/resource/Linear_transformation +
, http://dbpedia.org/resource/Positive-definite_matrix +
, http://dbpedia.org/resource/Inner_product +
, http://dbpedia.org/resource/Reproducing_kernel_Hilbert_space +
, http://dbpedia.org/resource/Time_series +
, http://dbpedia.org/resource/Feature_selection +
, http://dbpedia.org/resource/Cross-validation_%28statistics%29 +
, http://dbpedia.org/resource/Linear_map +
, http://dbpedia.org/resource/Kullback%E2%80%93Leibler_divergence +
|
| http://dbpedia.org/property/date
|
March 2020
|
| http://dbpedia.org/property/reason
|
This nonsense of calling a distribution P, with a capital X, when capital X is also the name of the random variable, and other like things, need to get cleaned up.
|
| http://dbpedia.org/property/wikiPageUsesTemplate
|
http://dbpedia.org/resource/Template:Reflist +
, http://dbpedia.org/resource/Template:Cleanup +
|
| http://purl.org/dc/terms/subject
|
http://dbpedia.org/resource/Category:Machine_learning +
, http://dbpedia.org/resource/Category:Theory_of_probability_distributions +
|
| http://www.w3.org/ns/prov#wasDerivedFrom
|
http://en.wikipedia.org/wiki/Kernel_embedding_of_distributions?oldid=1073073352&ns=0 +
|
| http://xmlns.com/foaf/0.1/isPrimaryTopicOf
|
http://en.wikipedia.org/wiki/Kernel_embedding_of_distributions +
|
| owl:sameAs |
https://global.dbpedia.org/id/ag3F +
, http://www.wikidata.org/entity/Q16000131 +
, http://dbpedia.org/resource/Kernel_embedding_of_distributions +
, http://rdf.freebase.com/ns/m.0zrsfd7 +
|
| rdfs:comment |
In machine learning, the kernel embedding … In machine learning, the kernel embedding of distributions (also called the kernel mean or mean map) comprises a class of nonparametric methods in which a probability distribution is represented as an element of a reproducing kernel Hilbert space (RKHS). A generalization of the individual data-point feature mapping done in classical kernel methods, the embedding of distributions into infinite-dimensional feature spaces can preserve all of the statistical features of arbitrary distributions, while allowing one to compare and manipulate distributions using Hilbert space operations such as inner products, distances, projections, linear transformations, and spectral analysis. This learning framework is very general and can be applied to distributions over any space on which a sensible kernel over any space on which a sensible kernel
|
| rdfs:label |
Kernel embedding of distributions
|
