Matches in SemOpenAlex for { <https://semopenalex.org/work/W3003921807> ?p ?o ?g. }
- W3003921807 abstract "We consider the problem of sampling from a strongly log-concave density in Rd, and prove an information theoretic lower bound on the number of stochastic gradient queries of the log density needed. Several popular sampling algorithms (including many Markov chain Monte Carlo methods) operate by using stochastic gradients of the log density to generate a sample; our results establish an information theoretic limit for all these algorithms. We show that for every algorithm, there exists a well-conditioned strongly log-concave target density for which the distribution of points generated by the algorithm would be at least ε away from the target in total variation distance if the number of gradient queries is less than Ω(σ2d/ε2), where σ2d is the variance of the stochastic gradient. Our lower bound follows by combining the ideas of Le Cam deficiency routinely used in the comparison of statistical experiments along with standard information theoretic tools used in lower bounding Bayes risk functions. To the best of our knowledge our results provide the first nontrivial dimension-dependent lower bound for this problem." @default.
- W3003921807 created "2020-02-07" @default.
- W3003921807 creator A5030261391 @default.
- W3003921807 creator A5033738047 @default.
- W3003921807 creator A5084487318 @default.
- W3003921807 date "2022-05-01" @default.
- W3003921807 modified "2023-09-24" @default.
- W3003921807 title "Oracle lower bounds for stochastic gradient sampling algorithms" @default.
- W3003921807 cites W116104437 @default.
- W3003921807 cites W1505731132 @default.
- W3003921807 cites W1505771370 @default.
- W3003921807 cites W1555683961 @default.
- W3003921807 cites W1558572410 @default.
- W3003921807 cites W1559116151 @default.
- W3003921807 cites W1971159979 @default.
- W3003921807 cites W1973594349 @default.
- W3003921807 cites W1983452151 @default.
- W3003921807 cites W1985093013 @default.
- W3003921807 cites W2017170340 @default.
- W3003921807 cites W2031715986 @default.
- W3003921807 cites W2042621213 @default.
- W3003921807 cites W2043657648 @default.
- W3003921807 cites W2050495532 @default.
- W3003921807 cites W2063342497 @default.
- W3003921807 cites W2063986634 @default.
- W3003921807 cites W2082399300 @default.
- W3003921807 cites W2082856185 @default.
- W3003921807 cites W2087273767 @default.
- W3003921807 cites W2096840748 @default.
- W3003921807 cites W2125299871 @default.
- W3003921807 cites W2155192981 @default.
- W3003921807 cites W2159000620 @default.
- W3003921807 cites W2167433878 @default.
- W3003921807 cites W2222154095 @default.
- W3003921807 cites W2478027467 @default.
- W3003921807 cites W2592172378 @default.
- W3003921807 cites W2746461338 @default.
- W3003921807 cites W2809641361 @default.
- W3003921807 cites W2883697110 @default.
- W3003921807 cites W2912099989 @default.
- W3003921807 cites W2962794482 @default.
- W3003921807 cites W2963599479 @default.
- W3003921807 cites W2963813262 @default.
- W3003921807 cites W2963964450 @default.
- W3003921807 cites W2964301034 @default.
- W3003921807 cites W2971331792 @default.
- W3003921807 cites W2997942410 @default.
- W3003921807 cites W3023788998 @default.
- W3003921807 cites W3034735417 @default.
- W3003921807 cites W3093571331 @default.
- W3003921807 cites W3147513489 @default.
- W3003921807 cites W61655770 @default.
- W3003921807 cites W621546036 @default.
- W3003921807 doi "https://doi.org/10.3150/21-bej1377" @default.
- W3003921807 hasPublicationYear "2022" @default.
- W3003921807 type Work @default.
- W3003921807 sameAs 3003921807 @default.
- W3003921807 citedByCount "3" @default.
- W3003921807 countsByYear W30039218072020 @default.
- W3003921807 countsByYear W30039218072021 @default.
- W3003921807 crossrefType "journal-article" @default.
- W3003921807 hasAuthorship W3003921807A5030261391 @default.
- W3003921807 hasAuthorship W3003921807A5033738047 @default.
- W3003921807 hasAuthorship W3003921807A5084487318 @default.
- W3003921807 hasBestOaLocation W30039218072 @default.
- W3003921807 hasConcept C105795698 @default.
- W3003921807 hasConcept C106131492 @default.
- W3003921807 hasConcept C111350023 @default.
- W3003921807 hasConcept C11413529 @default.
- W3003921807 hasConcept C114614502 @default.
- W3003921807 hasConcept C115903868 @default.
- W3003921807 hasConcept C134306372 @default.
- W3003921807 hasConcept C140779682 @default.
- W3003921807 hasConcept C154945302 @default.
- W3003921807 hasConcept C170593435 @default.
- W3003921807 hasConcept C19499675 @default.
- W3003921807 hasConcept C31972630 @default.
- W3003921807 hasConcept C33676613 @default.
- W3003921807 hasConcept C33923547 @default.
- W3003921807 hasConcept C41008148 @default.
- W3003921807 hasConcept C55166926 @default.
- W3003921807 hasConcept C62100291 @default.
- W3003921807 hasConcept C63584917 @default.
- W3003921807 hasConcept C77553402 @default.
- W3003921807 hasConcept C98763669 @default.
- W3003921807 hasConceptScore W3003921807C105795698 @default.
- W3003921807 hasConceptScore W3003921807C106131492 @default.
- W3003921807 hasConceptScore W3003921807C111350023 @default.
- W3003921807 hasConceptScore W3003921807C11413529 @default.
- W3003921807 hasConceptScore W3003921807C114614502 @default.
- W3003921807 hasConceptScore W3003921807C115903868 @default.
- W3003921807 hasConceptScore W3003921807C134306372 @default.
- W3003921807 hasConceptScore W3003921807C140779682 @default.
- W3003921807 hasConceptScore W3003921807C154945302 @default.
- W3003921807 hasConceptScore W3003921807C170593435 @default.
- W3003921807 hasConceptScore W3003921807C19499675 @default.
- W3003921807 hasConceptScore W3003921807C31972630 @default.
- W3003921807 hasConceptScore W3003921807C33676613 @default.
- W3003921807 hasConceptScore W3003921807C33923547 @default.
- W3003921807 hasConceptScore W3003921807C41008148 @default.