Matches in SemOpenAlex for { <https://semopenalex.org/work/W4301178504> ?p ?o ?g. }
- W4301178504 abstract "Restricted Boltzmann Machines are simple and powerful generative models that can encode any complex dataset. Despite all their advantages, in practice the trainings are often unstable and it is difficult to assess their quality because the dynamics are affected by extremely slow time dependencies. This situation becomes critical when dealing with low-dimensional clustered datasets, where the time required to sample ergodically the trained models becomes computationally prohibitive. In this work, we show that this divergence of Monte Carlo mixing times is related to a phenomenon of phase coexistence, similar to that which occurs in physics near a first-order phase transition. We show that sampling the equilibrium distribution using the Markov chain Monte Carlo method can be dramatically accelerated when using biased sampling techniques, in particular the Tethered Monte Carlo (TMC) method. This sampling technique efficiently solves the problem of evaluating the quality of a given trained model and generating new samples in a reasonable amount of time. Moreover, we show that this sampling technique can also be used to improve the computation of the log-likelihood gradient during training, leading to dramatic improvements in training RBMs with artificial clustered datasets. On real low-dimensional datasets, this new training method fits RBM models with significantly faster relaxation dynamics than those obtained with standard PCD recipes. We also show that TMC sampling can be used to recover the free-energy profile of the RBM. This proves to be extremely useful to compute the probability distribution of a given model and to improve the generation of new decorrelated samples in slow PCD-trained models." @default.
- W4301178504 created "2022-10-04" @default.
- W4301178504 creator A5001525923 @default.
- W4301178504 creator A5011613939 @default.
- W4301178504 creator A5057697784 @default.
- W4301178504 creator A5063165839 @default.
- W4301178504 date "2023-03-14" @default.
- W4301178504 modified "2023-10-15" @default.
- W4301178504 title "Learning a restricted Boltzmann machine using biased Monte Carlo sampling" @default.
- W4301178504 cites W1579788182 @default.
- W4301178504 cites W1971705019 @default.
- W4301178504 cites W1986142711 @default.
- W4301178504 cites W1991802452 @default.
- W4301178504 cites W1993177346 @default.
- W4301178504 cites W1996569672 @default.
- W4301178504 cites W2007339694 @default.
- W4301178504 cites W2031794121 @default.
- W4301178504 cites W2062375451 @default.
- W4301178504 cites W2100495367 @default.
- W4301178504 cites W2104549677 @default.
- W4301178504 cites W2116064496 @default.
- W4301178504 cites W2116825644 @default.
- W4301178504 cites W2136922672 @default.
- W4301178504 cites W2155458420 @default.
- W4301178504 cites W2556341799 @default.
- W4301178504 cites W2593807178 @default.
- W4301178504 cites W2769068614 @default.
- W4301178504 cites W2781744395 @default.
- W4301178504 cites W2914341699 @default.
- W4301178504 cites W2963640180 @default.
- W4301178504 cites W2973498290 @default.
- W4301178504 cites W2982859738 @default.
- W4301178504 cites W3096831136 @default.
- W4301178504 cites W3104390546 @default.
- W4301178504 cites W3104798141 @default.
- W4301178504 cites W3111171798 @default.
- W4301178504 cites W3127763315 @default.
- W4301178504 cites W3156867879 @default.
- W4301178504 cites W3164698142 @default.
- W4301178504 cites W3182036650 @default.
- W4301178504 cites W3204668690 @default.
- W4301178504 cites W4229002239 @default.
- W4301178504 cites W4229632203 @default.
- W4301178504 cites W4255825300 @default.
- W4301178504 cites W4296248562 @default.
- W4301178504 cites W4311415873 @default.
- W4301178504 cites W4316810617 @default.
- W4301178504 doi "https://doi.org/10.21468/scipostphys.14.3.032" @default.
- W4301178504 hasPublicationYear "2023" @default.
- W4301178504 type Work @default.
- W4301178504 citedByCount "1" @default.
- W4301178504 countsByYear W43011785042023 @default.
- W4301178504 crossrefType "journal-article" @default.
- W4301178504 hasAuthorship W4301178504A5001525923 @default.
- W4301178504 hasAuthorship W4301178504A5011613939 @default.
- W4301178504 hasAuthorship W4301178504A5057697784 @default.
- W4301178504 hasAuthorship W4301178504A5063165839 @default.
- W4301178504 hasBestOaLocation W43011785041 @default.
- W4301178504 hasConcept C105795698 @default.
- W4301178504 hasConcept C106131492 @default.
- W4301178504 hasConcept C107673813 @default.
- W4301178504 hasConcept C111350023 @default.
- W4301178504 hasConcept C11413529 @default.
- W4301178504 hasConcept C119857082 @default.
- W4301178504 hasConcept C121332964 @default.
- W4301178504 hasConcept C121864883 @default.
- W4301178504 hasConcept C13153151 @default.
- W4301178504 hasConcept C132725507 @default.
- W4301178504 hasConcept C138885662 @default.
- W4301178504 hasConcept C140779682 @default.
- W4301178504 hasConcept C154945302 @default.
- W4301178504 hasConcept C187192777 @default.
- W4301178504 hasConcept C192576344 @default.
- W4301178504 hasConcept C19499675 @default.
- W4301178504 hasConcept C207390915 @default.
- W4301178504 hasConcept C31972630 @default.
- W4301178504 hasConcept C33923547 @default.
- W4301178504 hasConcept C41008148 @default.
- W4301178504 hasConcept C41895202 @default.
- W4301178504 hasConcept C50644808 @default.
- W4301178504 hasConcept C52740198 @default.
- W4301178504 hasConcept C98763669 @default.
- W4301178504 hasConceptScore W4301178504C105795698 @default.
- W4301178504 hasConceptScore W4301178504C106131492 @default.
- W4301178504 hasConceptScore W4301178504C107673813 @default.
- W4301178504 hasConceptScore W4301178504C111350023 @default.
- W4301178504 hasConceptScore W4301178504C11413529 @default.
- W4301178504 hasConceptScore W4301178504C119857082 @default.
- W4301178504 hasConceptScore W4301178504C121332964 @default.
- W4301178504 hasConceptScore W4301178504C121864883 @default.
- W4301178504 hasConceptScore W4301178504C13153151 @default.
- W4301178504 hasConceptScore W4301178504C132725507 @default.
- W4301178504 hasConceptScore W4301178504C138885662 @default.
- W4301178504 hasConceptScore W4301178504C140779682 @default.
- W4301178504 hasConceptScore W4301178504C154945302 @default.
- W4301178504 hasConceptScore W4301178504C187192777 @default.
- W4301178504 hasConceptScore W4301178504C192576344 @default.
- W4301178504 hasConceptScore W4301178504C19499675 @default.
- W4301178504 hasConceptScore W4301178504C207390915 @default.
- W4301178504 hasConceptScore W4301178504C31972630 @default.