Matches in SemOpenAlex for { <https://semopenalex.org/work/W3182036650> ?p ?o ?g. }
- W3182036650 abstract "Restricted Boltzmann machines (RBM) are bilayer neural networks used for the unsupervised learning of model distributions from data. The bipartite architecture of RBM naturally defines an elegant sampling procedure, called alternating Gibbs sampling (AGS), where the configurations of the latent-variable layer are sampled conditional to the data-variable layer and vice versa. We study here the performance of AGS on several analytically tractable models borrowed from statistical mechanics. We show that standard AGS is not more efficient than classical Metropolis-Hastings (MH) sampling of the effective energy landscape defined on the data layer. However, RBM can identify meaningful representations of training data in their latent space. Furthermore, using these representations and combining Gibbs sampling with the MH algorithm in the latent space can enhance the sampling performance of the RBM when the hidden units encode weakly dependent features of the data. We illustrate our findings on three datasets: Bars and Stripes and MNIST, well known in machine learning, and the so-called lattice proteins dataset, introduced in theoretical biology to study the sequence-to-structure mapping in proteins." @default.
- W3182036650 created "2021-07-19" @default.
- W3182036650 creator A5016002177 @default.
- W3182036650 creator A5050420323 @default.
- W3182036650 creator A5070078482 @default.
- W3182036650 date "2021-09-07" @default.
- W3182036650 modified "2023-10-11" @default.
- W3182036650 title "Barriers and dynamical paths in alternating Gibbs sampling of restricted Boltzmann machines" @default.
- W3182036650 cites W1884390770 @default.
- W3182036650 cites W1969758109 @default.
- W3182036650 cites W1983272724 @default.
- W3182036650 cites W2020999234 @default.
- W3182036650 cites W2030358500 @default.
- W3182036650 cites W2037139490 @default.
- W3182036650 cites W2051203581 @default.
- W3182036650 cites W2064630666 @default.
- W3182036650 cites W2069129925 @default.
- W3182036650 cites W2070143079 @default.
- W3182036650 cites W2071108817 @default.
- W3182036650 cites W2072761075 @default.
- W3182036650 cites W2076758466 @default.
- W3182036650 cites W2116064496 @default.
- W3182036650 cites W2120272360 @default.
- W3182036650 cites W2125150881 @default.
- W3182036650 cites W2128084896 @default.
- W3182036650 cites W2130949317 @default.
- W3182036650 cites W2136922672 @default.
- W3182036650 cites W2138309709 @default.
- W3182036650 cites W2154335411 @default.
- W3182036650 cites W2171546617 @default.
- W3182036650 cites W2391698393 @default.
- W3182036650 cites W2530117613 @default.
- W3182036650 cites W2530819665 @default.
- W3182036650 cites W2556341799 @default.
- W3182036650 cites W2593807178 @default.
- W3182036650 cites W2615473567 @default.
- W3182036650 cites W2736723884 @default.
- W3182036650 cites W2756015143 @default.
- W3182036650 cites W2769068614 @default.
- W3182036650 cites W2781744395 @default.
- W3182036650 cites W2781887058 @default.
- W3182036650 cites W2792312473 @default.
- W3182036650 cites W2883782791 @default.
- W3182036650 cites W2905410813 @default.
- W3182036650 cites W2911964244 @default.
- W3182036650 cites W2963640180 @default.
- W3182036650 cites W3156867879 @default.
- W3182036650 cites W3173484311 @default.
- W3182036650 cites W4238020138 @default.
- W3182036650 doi "https://doi.org/10.1103/physreve.104.034109" @default.
- W3182036650 hasPubMedId "https://pubmed.ncbi.nlm.nih.gov/34654094" @default.
- W3182036650 hasPublicationYear "2021" @default.
- W3182036650 type Work @default.
- W3182036650 sameAs 3182036650 @default.
- W3182036650 citedByCount "9" @default.
- W3182036650 countsByYear W31820366502021 @default.
- W3182036650 countsByYear W31820366502022 @default.
- W3182036650 countsByYear W31820366502023 @default.
- W3182036650 crossrefType "journal-article" @default.
- W3182036650 hasAuthorship W3182036650A5016002177 @default.
- W3182036650 hasAuthorship W3182036650A5050420323 @default.
- W3182036650 hasAuthorship W3182036650A5070078482 @default.
- W3182036650 hasBestOaLocation W31820366502 @default.
- W3182036650 hasConcept C106131492 @default.
- W3182036650 hasConcept C107673813 @default.
- W3182036650 hasConcept C111919701 @default.
- W3182036650 hasConcept C121332964 @default.
- W3182036650 hasConcept C121864883 @default.
- W3182036650 hasConcept C140779682 @default.
- W3182036650 hasConcept C154945302 @default.
- W3182036650 hasConcept C158424031 @default.
- W3182036650 hasConcept C190502265 @default.
- W3182036650 hasConcept C192576344 @default.
- W3182036650 hasConcept C193999330 @default.
- W3182036650 hasConcept C199354608 @default.
- W3182036650 hasConcept C2778572836 @default.
- W3182036650 hasConcept C31972630 @default.
- W3182036650 hasConcept C41008148 @default.
- W3182036650 hasConcept C50644808 @default.
- W3182036650 hasConcept C51167844 @default.
- W3182036650 hasConcept C80444323 @default.
- W3182036650 hasConceptScore W3182036650C106131492 @default.
- W3182036650 hasConceptScore W3182036650C107673813 @default.
- W3182036650 hasConceptScore W3182036650C111919701 @default.
- W3182036650 hasConceptScore W3182036650C121332964 @default.
- W3182036650 hasConceptScore W3182036650C121864883 @default.
- W3182036650 hasConceptScore W3182036650C140779682 @default.
- W3182036650 hasConceptScore W3182036650C154945302 @default.
- W3182036650 hasConceptScore W3182036650C158424031 @default.
- W3182036650 hasConceptScore W3182036650C190502265 @default.
- W3182036650 hasConceptScore W3182036650C192576344 @default.
- W3182036650 hasConceptScore W3182036650C193999330 @default.
- W3182036650 hasConceptScore W3182036650C199354608 @default.
- W3182036650 hasConceptScore W3182036650C2778572836 @default.
- W3182036650 hasConceptScore W3182036650C31972630 @default.
- W3182036650 hasConceptScore W3182036650C41008148 @default.
- W3182036650 hasConceptScore W3182036650C50644808 @default.
- W3182036650 hasConceptScore W3182036650C51167844 @default.
- W3182036650 hasConceptScore W3182036650C80444323 @default.
- W3182036650 hasFunder F4320320883 @default.