Matches in SemOpenAlex for { <https://semopenalex.org/work/W3003552232> ?p ?o ?g. }
- W3003552232 abstract "In this work, we focus on variational Bayesian inference on the sparse Deep Neural Network (DNN) modeled under a class of spike-and-slab priors. Given a pre-specified sparse DNN structure, the corresponding variational posterior contraction rate is characterized that reveals a trade-off between the variational error and the approximation error, which are both determined by the network structural complexity (i.e., depth, width and sparsity). However, the optimal network structure, which strikes the balance of the aforementioned trade-off and yields the best rate, is generally unknown in reality. Therefore, our work further develops an {em adaptive} variational inference procedure that can automatically select a reasonably good (data-dependent) network structure that achieves the best contraction rate, without knowing the optimal network structure. In particular, when the true function is H{o}lder smooth, the adaptive variational inference is capable to attain (near-)optimal rate without the knowledge of smoothness level. The above rate still suffers from the curse of dimensionality, and thus motivates the teacher-student setup, i.e., the true function is a sparse DNN model, under which the rate only logarithmically depends on the input dimension." @default.
- W3003552232 created "2020-02-07" @default.
- W3003552232 creator A5005876602 @default.
- W3003552232 creator A5042490620 @default.
- W3003552232 creator A5075942874 @default.
- W3003552232 date "2019-10-10" @default.
- W3003552232 modified "2023-09-27" @default.
- W3003552232 title "Adaptive Variational Bayesian Inference for Sparse Deep Neural Network" @default.
- W3003552232 cites W1516111018 @default.
- W3003552232 cites W1535258871 @default.
- W3003552232 cites W1719489212 @default.
- W3003552232 cites W1968562765 @default.
- W3003552232 cites W2007069447 @default.
- W3003552232 cites W2064381290 @default.
- W3003552232 cites W2095705004 @default.
- W3003552232 cites W2103496339 @default.
- W3003552232 cites W2108677974 @default.
- W3003552232 cites W2111051539 @default.
- W3003552232 cites W2149991584 @default.
- W3003552232 cites W2156387975 @default.
- W3003552232 cites W2164411961 @default.
- W3003552232 cites W2547875792 @default.
- W3003552232 cites W2548228487 @default.
- W3003552232 cites W2557283755 @default.
- W3003552232 cites W2582745083 @default.
- W3003552232 cites W2604268533 @default.
- W3003552232 cites W2619650475 @default.
- W3003552232 cites W2726058692 @default.
- W3003552232 cites W2763388430 @default.
- W3003552232 cites W2776889781 @default.
- W3003552232 cites W2794607344 @default.
- W3003552232 cites W2893242978 @default.
- W3003552232 cites W2915589364 @default.
- W3003552232 cites W2946302218 @default.
- W3003552232 cites W2951004968 @default.
- W3003552232 cites W2951266961 @default.
- W3003552232 cites W2962761333 @default.
- W3003552232 cites W2962897886 @default.
- W3003552232 cites W2963131860 @default.
- W3003552232 cites W2963243833 @default.
- W3003552232 cites W2964191999 @default.
- W3003552232 cites W2970540478 @default.
- W3003552232 cites W2971678291 @default.
- W3003552232 cites W3049347563 @default.
- W3003552232 cites W3104051756 @default.
- W3003552232 cites W568673721 @default.
- W3003552232 hasPublicationYear "2019" @default.
- W3003552232 type Work @default.
- W3003552232 sameAs 3003552232 @default.
- W3003552232 citedByCount "0" @default.
- W3003552232 crossrefType "posted-content" @default.
- W3003552232 hasAuthorship W3003552232A5005876602 @default.
- W3003552232 hasAuthorship W3003552232A5042490620 @default.
- W3003552232 hasAuthorship W3003552232A5075942874 @default.
- W3003552232 hasConcept C102634674 @default.
- W3003552232 hasConcept C107673813 @default.
- W3003552232 hasConcept C111030470 @default.
- W3003552232 hasConcept C11413529 @default.
- W3003552232 hasConcept C126255220 @default.
- W3003552232 hasConcept C134306372 @default.
- W3003552232 hasConcept C154945302 @default.
- W3003552232 hasConcept C160234255 @default.
- W3003552232 hasConcept C177769412 @default.
- W3003552232 hasConcept C2776214188 @default.
- W3003552232 hasConcept C33923547 @default.
- W3003552232 hasConcept C40969351 @default.
- W3003552232 hasConcept C41008148 @default.
- W3003552232 hasConcept C50644808 @default.
- W3003552232 hasConceptScore W3003552232C102634674 @default.
- W3003552232 hasConceptScore W3003552232C107673813 @default.
- W3003552232 hasConceptScore W3003552232C111030470 @default.
- W3003552232 hasConceptScore W3003552232C11413529 @default.
- W3003552232 hasConceptScore W3003552232C126255220 @default.
- W3003552232 hasConceptScore W3003552232C134306372 @default.
- W3003552232 hasConceptScore W3003552232C154945302 @default.
- W3003552232 hasConceptScore W3003552232C160234255 @default.
- W3003552232 hasConceptScore W3003552232C177769412 @default.
- W3003552232 hasConceptScore W3003552232C2776214188 @default.
- W3003552232 hasConceptScore W3003552232C33923547 @default.
- W3003552232 hasConceptScore W3003552232C40969351 @default.
- W3003552232 hasConceptScore W3003552232C41008148 @default.
- W3003552232 hasConceptScore W3003552232C50644808 @default.
- W3003552232 hasLocation W30035522321 @default.
- W3003552232 hasOpenAccess W3003552232 @default.
- W3003552232 hasPrimaryLocation W30035522321 @default.
- W3003552232 hasRelatedWork W1504002966 @default.
- W3003552232 hasRelatedWork W1512149552 @default.
- W3003552232 hasRelatedWork W1562453744 @default.
- W3003552232 hasRelatedWork W1790658179 @default.
- W3003552232 hasRelatedWork W2115120746 @default.
- W3003552232 hasRelatedWork W2276728651 @default.
- W3003552232 hasRelatedWork W2292972218 @default.
- W3003552232 hasRelatedWork W2608781926 @default.
- W3003552232 hasRelatedWork W2766159815 @default.
- W3003552232 hasRelatedWork W2770888898 @default.
- W3003552232 hasRelatedWork W2898043227 @default.
- W3003552232 hasRelatedWork W2928453536 @default.
- W3003552232 hasRelatedWork W2941256613 @default.
- W3003552232 hasRelatedWork W3008547934 @default.
- W3003552232 hasRelatedWork W3106331161 @default.