This HTML5 document contains 119 embedded RDF statements represented using HTML+Microdata notation.

The embedded RDF content will be recognized by any processor of HTML5 Microdata.

Namespace Prefixes

PrefixIRI
n6http://www.csl.sony.co.jp/person/nielsen/PT/SoCG07/
dctermshttp://purl.org/dc/terms/
yago-reshttp://yago-knowledge.org/resource/
dbohttp://dbpedia.org/ontology/
foafhttp://xmlns.com/foaf/0.1/
n5http://dbpedia.org/resource/File:
n23https://web.archive.org/web/20100812221422/http:/www.ee.washington.edu/research/guptalab/publications/
n21http://www.ee.washington.edu/research/guptalab/publications/
n24https://global.dbpedia.org/id/
yagohttp://dbpedia.org/class/yago/
dbpedia-ruhttp://ru.dbpedia.org/resource/
dbthttp://dbpedia.org/resource/Template:
rdfshttp://www.w3.org/2000/01/rdf-schema#
freebasehttp://rdf.freebase.com/ns/
n14http://commons.wikimedia.org/wiki/Special:FilePath/
rdfhttp://www.w3.org/1999/02/22-rdf-syntax-ns#
owlhttp://www.w3.org/2002/07/owl#
n16https://web.archive.org/web/20170217025324/https:/www2.ee.washington.edu/techsite/papers/documents/
n28https://www.ee.washington.edu/techsite/papers/documents/
n26http://jmlr.csail.mit.edu/papers/v6/
wikipedia-enhttp://en.wikipedia.org/wiki/
dbpedia-frhttp://fr.dbpedia.org/resource/
n8http://hal.archives-ouvertes.fr/hal-00488441/en/
dbphttp://dbpedia.org/property/
dbchttp://dbpedia.org/resource/Category:
provhttp://www.w3.org/ns/prov#
n29http://www.lix.polytechnique.fr/~nielsen/
xsdhhttp://www.w3.org/2001/XMLSchema#
wikidatahttp://www.wikidata.org/entity/
dbrhttp://dbpedia.org/resource/

Statements

Subject Item
dbr:Bregman_divergence
rdf:type
yago:Act100030358 yago:WikicatGeometricAlgorithms yago:Choice100161243 yago:PsychologicalFeature100023100 yago:Maneuver100168237 yago:WikicatStatisticalDistanceMeasures yago:Action100037396 yago:Move100165942 yago:Rule105846932 yago:Abstraction100002137 yago:Measure100174412 yago:Decision100162632 yago:Event100029378 yago:Procedure101023820 yago:Algorithm105847438 yago:Activity100407535 yago:YagoPermanentlyLocatedEntity
rdfs:label
Дивергенция Брэгмана Divergence de Bregman Bregman divergence
rdfs:comment
In mathematics, specifically statistics and information geometry, a Bregman divergence or Bregman distance is a measure of difference between two points, defined in terms of a strictly convex function; they form an important class of divergences. When the points are interpreted as probability distributions – notably as either values of the parameter of a parametric model or as a data set of observed values – the resulting distance is a statistical distance. The most basic Bregman divergence is the squared Euclidean distance. Дивергенция Брэгмана или расстояние Брэгмана — мера расстояния между двумя точками, определённая в терминах строго выпуклой функции. Они образуют важный класс дивергенций. Если точки интерпретировать как распределение вероятностей, либо как значения , либо как набор наблюдаемых значений, то полученное расстояние является . Самой элементарной дивергенцией Брэгмана является квадрат евклидова расстояния. Дивергенция Брэгмана названа по имени Льва Мееровича Брэгмана, предложившего концепцию в 1967 году. En mathématiques, la divergence de Bregman est une mesure de la différence entre deux distributions dérivée d'une fonction potentiel U à valeurs réelles strictement convexe et continûment différentiable. Le concept a été introduit par (en) en 1967.Par l'intermédiaire de la transformation de Legendre, au potentiel correspond un potentiel dual et leur différentiation donne naissance à deux systèmes de coordonnées duaux.
foaf:depiction
n14:Bregman_divergence_diagram_used_in_proof_of_squared_generalized_Euclidean_distances.png n14:Bregman_divergence_Pythagorean.png
dcterms:subject
dbc:Statistical_distance dbc:Geometric_algorithms
dbo:wikiPageID
4491248
dbo:wikiPageRevisionID
1099751081
dbo:wikiPageWikiLink
dbr:Least_squares dbr:Information_geometry n5:Bregman_divergence_Pythagorean.png n5:Bregman_divergence_diagram_used_in_proof_of_squared_generalized_Euclidean_distances.png dbr:Mutual_information dbr:Delaunay_triangulation dbr:Mahalanobis_distance dbr:Data_processing_inequality dbr:Squared_Euclidean_distance dbr:Kullback–Leibler_divergence dbr:Computational_geometry dbr:Conference_on_Neural_Information_Processing_Systems dbr:Statistical_manifold dbr:Lev_M._Bregman dbr:Von_Neumann_entropy dbr:Hamming_distance dbr:Jensen's_inequality dbr:Taylor_expansion dbr:Parametric_model dbr:Statistics dbr:Precision_and_recall dbr:Mirror_descent dbr:Probability_distribution dbr:Convex_conjugate dbc:Statistical_distance dbr:Projective_duality dbr:Convex_function dbr:Pythagorean_theorem dbr:Gradient_descent dbr:Divergence_(statistics) dbr:F-divergence dbr:Mathematics dbr:Optimization_theory dbr:IEEE_Transactions_on_Information_Theory dbr:Statistical_distance dbr:Journal_of_Machine_Learning_Research dbr:Flat_manifold dbr:Stein's_loss dbr:Itakura–Saito_distance dbr:Entropy_(information_theory) dbr:Positive_definiteness dbr:Convex_set dbr:Parallelogram_law dbr:Metric_(mathematics) dbr:Relative_interior dbr:Softmax_function dbr:Hedge_algorithm dbr:Discrete_and_Computational_Geometry dbc:Geometric_algorithms dbr:Voronoi_diagram dbr:IEEE_Signal_Processing_Letters dbr:Jensen–Shannon_divergence dbr:Submodular_set_function dbr:Triangle_inequality
dbo:wikiPageExternalLink
n6: n8: n16:UWEETR-2008-0001.pdf n21:FrigyikSrivastavaGupta.pdf n23:FrigyikSrivastavaGupta.pdf n26:banerjee05b.html n28:UWEETR-2008-0001.pdf n29:ISVD09-GenBregmanVD.pdf
owl:sameAs
yago-res:Bregman_divergence dbpedia-fr:Divergence_de_Bregman wikidata:Q4960382 freebase:m.0c5b8d n24:4be6d dbpedia-ru:Дивергенция_Брэгмана
dbp:wikiPageUsesTemplate
dbt:Dead_link dbt:Math_proof dbt:Short_description dbt:Cite_arXiv dbt:Cite_conference dbt:Refbegin dbt:Cite_book dbt:Cite_journal dbt:Refend dbt:Reflist
dbo:thumbnail
n14:Bregman_divergence_Pythagorean.png?width=300
dbp:bot
InternetArchiveBot
dbp:date
June 2019
dbp:fixAttempted
yes
dbp:proof
thumb|Bregman divergence interpreted as areas. For any , define for . Let . Then for , and since is continuous, also for . Then, from the diagram, we see that for for all , we must have linear on . Thus we find that varies linearly along any direction. By the next lemma, is quadratic. Since is also strictly convex, it is of form , where . Lemma: If is an open subset of , has continuous derivative, and given any line segment , the function is linear in , then is a quadratic function. Proof idea: For any quadratic function , we have still has such derivative-linearity, so we will subtract away a few quadratic functions and show that becomes zero. The proof idea can be illustrated fully for the case of , so we prove it in this case. By the derivative-linearity, is a quadratic function on any line segment in . We subtract away four quadratic functions, such that becomes identically zero on the x-axis, y-axis, and the line. Let , for well-chosen . Now use to remove the linear term, and use respectively to remove the quadratic terms along the three lines. not on the origin, there exists a line across that intersects the x-axis, y-axis, and the line at three different points. Since is quadratic on , and is zero on three different points, is identically zero on , thus . Thus is quadratic.
dbp:drop
hidden
dbo:abstract
In mathematics, specifically statistics and information geometry, a Bregman divergence or Bregman distance is a measure of difference between two points, defined in terms of a strictly convex function; they form an important class of divergences. When the points are interpreted as probability distributions – notably as either values of the parameter of a parametric model or as a data set of observed values – the resulting distance is a statistical distance. The most basic Bregman divergence is the squared Euclidean distance. Bregman divergences are similar to metrics, but satisfy neither the triangle inequality (ever) nor symmetry (in general). However, they satisfy a generalization of the Pythagorean theorem, and in information geometry the corresponding statistical manifold is interpreted as a (dually) flat manifold. This allows many techniques of optimization theory to be generalized to Bregman divergences, geometrically as generalizations of least squares. Bregman divergences are named after Russian mathematician Lev M. Bregman, who introduced the concept in 1967. En mathématiques, la divergence de Bregman est une mesure de la différence entre deux distributions dérivée d'une fonction potentiel U à valeurs réelles strictement convexe et continûment différentiable. Le concept a été introduit par (en) en 1967.Par l'intermédiaire de la transformation de Legendre, au potentiel correspond un potentiel dual et leur différentiation donne naissance à deux systèmes de coordonnées duaux. Дивергенция Брэгмана или расстояние Брэгмана — мера расстояния между двумя точками, определённая в терминах строго выпуклой функции. Они образуют важный класс дивергенций. Если точки интерпретировать как распределение вероятностей, либо как значения , либо как набор наблюдаемых значений, то полученное расстояние является . Самой элементарной дивергенцией Брэгмана является квадрат евклидова расстояния. Дивергенции Брэгмана подобны метрикам, но не удовлетворяют ни неравенству треугольника, ни симметрии (в общем случае), однако они удовлетворяют обобщённой теореме Пифагора. В соответствующее интерпретируется как (или двойственное). Это позволяет обобщить многие техники оптимизации к дивергенции Брэгмана, что геометрически соответствует обобщению метода наименьших квадратов. Дивергенция Брэгмана названа по имени Льва Мееровича Брэгмана, предложившего концепцию в 1967 году.
prov:wasDerivedFrom
wikipedia-en:Bregman_divergence?oldid=1099751081&ns=0
dbo:wikiPageLength
24956
foaf:isPrimaryTopicOf
wikipedia-en:Bregman_divergence