Matches in SemOpenAlex for { <https://semopenalex.org/work/W3090222223> ?p ?o ?g. }
Showing items 1 to 95 of
95
with 100 items per page.
- W3090222223 abstract "Large scale Bayesian nonparametrics (BNP) learner such as Stochastic Variational Inference (SVI) can handle datasets with large class number and large training size at fractional cost. Like its predecessor, SVI rely on the assumption of conjugate variational posterior to approximate the true posterior. A more challenging problem is to consider large scale learning on non-conjugate posterior. Recent works in this direction are mostly associated with using Monte Carlo methods for approximating the learner. However, these works are usually demonstrated on non-BNP related task and less complex models such as logistic regression, due to higher computational complexity. In order to overcome the issue faced by SVI, we develop a novel approach based on the recently proposed constant stepsize stochastic gradient ascent to allow large scale learning on non-conjugate posterior. Unlike SVI, our new learner does not require closed-form expression for the variational posterior expectatations. Our only requirement is that the variational posterior is differentiable. In order to ensure convergence in stochastic settings, SVI rely on decaying step-sizes to slow its learning. Inspired by SVI and Adam, we propose the novel use of adaptive stepsizes in our method to significantly improve its learning. We show that our proposed methods is compatible with ResNet features when applied to large class number datasets such as MIT67 and SUN397. Finally, we compare our proposed learner with several recent works such as deep clustering algorithms and showed we were able to produce on-par or outperform the state-of-the-art methods in terms of clustering measures." @default.
- W3090222223 created "2020-10-08" @default.
- W3090222223 creator A5087657164 @default.
- W3090222223 date "2020-07-01" @default.
- W3090222223 modified "2023-09-27" @default.
- W3090222223 title "Non-conjugate Posterior using Stochastic Gradient Ascent with Adaptive Stepsize" @default.
- W3090222223 cites W1499798934 @default.
- W3090222223 cites W1663973292 @default.
- W3090222223 cites W166614460 @default.
- W3090222223 cites W1845773340 @default.
- W3090222223 cites W1880262756 @default.
- W3090222223 cites W1919588359 @default.
- W3090222223 cites W1970789124 @default.
- W3090222223 cites W1994616650 @default.
- W3090222223 cites W2005564522 @default.
- W3090222223 cites W2099471712 @default.
- W3090222223 cites W2119196781 @default.
- W3090222223 cites W2127498532 @default.
- W3090222223 cites W2163605009 @default.
- W3090222223 cites W2166851633 @default.
- W3090222223 cites W2167433878 @default.
- W3090222223 cites W2194775991 @default.
- W3090222223 cites W2413356069 @default.
- W3090222223 cites W2603986758 @default.
- W3090222223 cites W2767734436 @default.
- W3090222223 cites W2777922598 @default.
- W3090222223 cites W2798534672 @default.
- W3090222223 cites W2883725317 @default.
- W3090222223 cites W2948398419 @default.
- W3090222223 cites W2962761061 @default.
- W3090222223 cites W2962915600 @default.
- W3090222223 cites W2962994101 @default.
- W3090222223 cites W2963090522 @default.
- W3090222223 cites W2963173382 @default.
- W3090222223 cites W2964074409 @default.
- W3090222223 cites W3101380508 @default.
- W3090222223 cites W82771173 @default.
- W3090222223 doi "https://doi.org/10.1109/ijcnn48605.2020.9207115" @default.
- W3090222223 hasPublicationYear "2020" @default.
- W3090222223 type Work @default.
- W3090222223 sameAs 3090222223 @default.
- W3090222223 citedByCount "0" @default.
- W3090222223 crossrefType "proceedings-article" @default.
- W3090222223 hasAuthorship W3090222223A5087657164 @default.
- W3090222223 hasConcept C107673813 @default.
- W3090222223 hasConcept C11413529 @default.
- W3090222223 hasConcept C121332964 @default.
- W3090222223 hasConcept C126255220 @default.
- W3090222223 hasConcept C153258448 @default.
- W3090222223 hasConcept C154945302 @default.
- W3090222223 hasConcept C199360897 @default.
- W3090222223 hasConcept C206688291 @default.
- W3090222223 hasConcept C2777027219 @default.
- W3090222223 hasConcept C2778755073 @default.
- W3090222223 hasConcept C33923547 @default.
- W3090222223 hasConcept C41008148 @default.
- W3090222223 hasConcept C50644808 @default.
- W3090222223 hasConcept C57830394 @default.
- W3090222223 hasConcept C62520636 @default.
- W3090222223 hasConcept C73555534 @default.
- W3090222223 hasConcept C81184566 @default.
- W3090222223 hasConceptScore W3090222223C107673813 @default.
- W3090222223 hasConceptScore W3090222223C11413529 @default.
- W3090222223 hasConceptScore W3090222223C121332964 @default.
- W3090222223 hasConceptScore W3090222223C126255220 @default.
- W3090222223 hasConceptScore W3090222223C153258448 @default.
- W3090222223 hasConceptScore W3090222223C154945302 @default.
- W3090222223 hasConceptScore W3090222223C199360897 @default.
- W3090222223 hasConceptScore W3090222223C206688291 @default.
- W3090222223 hasConceptScore W3090222223C2777027219 @default.
- W3090222223 hasConceptScore W3090222223C2778755073 @default.
- W3090222223 hasConceptScore W3090222223C33923547 @default.
- W3090222223 hasConceptScore W3090222223C41008148 @default.
- W3090222223 hasConceptScore W3090222223C50644808 @default.
- W3090222223 hasConceptScore W3090222223C57830394 @default.
- W3090222223 hasConceptScore W3090222223C62520636 @default.
- W3090222223 hasConceptScore W3090222223C73555534 @default.
- W3090222223 hasConceptScore W3090222223C81184566 @default.
- W3090222223 hasLocation W30902222231 @default.
- W3090222223 hasOpenAccess W3090222223 @default.
- W3090222223 hasPrimaryLocation W30902222231 @default.
- W3090222223 hasRelatedWork W11353818 @default.
- W3090222223 hasRelatedWork W11910490 @default.
- W3090222223 hasRelatedWork W14625021 @default.
- W3090222223 hasRelatedWork W2700343 @default.
- W3090222223 hasRelatedWork W6908809 @default.
- W3090222223 hasRelatedWork W7439820 @default.
- W3090222223 hasRelatedWork W819137 @default.
- W3090222223 hasRelatedWork W8208422 @default.
- W3090222223 hasRelatedWork W8936527 @default.
- W3090222223 hasRelatedWork W13691832 @default.
- W3090222223 isParatext "false" @default.
- W3090222223 isRetracted "false" @default.
- W3090222223 magId "3090222223" @default.
- W3090222223 workType "article" @default.