Matches in SemOpenAlex for { <https://semopenalex.org/work/W2897834873> ?p ?o ?g. }
Showing items 1 to 83 of
83
with 100 items per page.
- W2897834873 abstract "Despite their impressive performance, Deep Neural Networks (DNNs) typically underperform Gradient Boosting Trees (GBTs) on many tabular-dataset learning tasks. We propose that applying a different regularization coefficient to each weight might boost the performance of DNNs by allowing them to make more use of the more relevant inputs. However, this will lead to an intractable number of hyperparameters. Here, we introduce Regularization Learning Networks (RLNs), which overcome this challenge by introducing an efficient hyperparameter tuning scheme which minimizes a new Counterfactual Loss. Our results show that RLNs significantly improve DNNs on tabular datasets, and achieve comparable results to GBTs, with the best performance achieved with an ensemble that combines GBTs and RLNs. RLNs produce extremely sparse networks, eliminating up to 99.8% of the network edges and 82% of the input features, thus providing more interpretable models and reveal the importance that the network assigns to different inputs. RLNs could efficiently learn a single network in datasets that comprise both tabular and unstructured data, such as in the setting of medical imaging accompanied by electronic health records. An open source implementation of RLN can be found at this https URL." @default.
- W2897834873 created "2018-10-26" @default.
- W2897834873 creator A5012450539 @default.
- W2897834873 creator A5081284206 @default.
- W2897834873 date "2018-05-16" @default.
- W2897834873 modified "2023-09-27" @default.
- W2897834873 title "Regularization Learning Networks: Deep Learning for Tabular Datasets" @default.
- W2897834873 cites W1833005471 @default.
- W2897834873 cites W1868018859 @default.
- W2897834873 cites W1945616565 @default.
- W2897834873 cites W2346578521 @default.
- W2897834873 cites W2404901863 @default.
- W2897834873 cites W2493343568 @default.
- W2897834873 cites W2731899572 @default.
- W2897834873 cites W2786022758 @default.
- W2897834873 cites W2795247881 @default.
- W2897834873 cites W2950182411 @default.
- W2897834873 cites W2951184134 @default.
- W2897834873 cites W2952111767 @default.
- W2897834873 cites W3098949126 @default.
- W2897834873 cites W3102476541 @default.
- W2897834873 cites W3183459807 @default.
- W2897834873 cites W60686164 @default.
- W2897834873 hasPublicationYear "2018" @default.
- W2897834873 type Work @default.
- W2897834873 sameAs 2897834873 @default.
- W2897834873 citedByCount "2" @default.
- W2897834873 countsByYear W28978348732019 @default.
- W2897834873 countsByYear W28978348732021 @default.
- W2897834873 crossrefType "posted-content" @default.
- W2897834873 hasAuthorship W2897834873A5012450539 @default.
- W2897834873 hasAuthorship W2897834873A5081284206 @default.
- W2897834873 hasConcept C108583219 @default.
- W2897834873 hasConcept C108650721 @default.
- W2897834873 hasConcept C111472728 @default.
- W2897834873 hasConcept C119857082 @default.
- W2897834873 hasConcept C124101348 @default.
- W2897834873 hasConcept C138885662 @default.
- W2897834873 hasConcept C154945302 @default.
- W2897834873 hasConcept C2776135515 @default.
- W2897834873 hasConcept C2984842247 @default.
- W2897834873 hasConcept C41008148 @default.
- W2897834873 hasConcept C50644808 @default.
- W2897834873 hasConcept C8642999 @default.
- W2897834873 hasConceptScore W2897834873C108583219 @default.
- W2897834873 hasConceptScore W2897834873C108650721 @default.
- W2897834873 hasConceptScore W2897834873C111472728 @default.
- W2897834873 hasConceptScore W2897834873C119857082 @default.
- W2897834873 hasConceptScore W2897834873C124101348 @default.
- W2897834873 hasConceptScore W2897834873C138885662 @default.
- W2897834873 hasConceptScore W2897834873C154945302 @default.
- W2897834873 hasConceptScore W2897834873C2776135515 @default.
- W2897834873 hasConceptScore W2897834873C2984842247 @default.
- W2897834873 hasConceptScore W2897834873C41008148 @default.
- W2897834873 hasConceptScore W2897834873C50644808 @default.
- W2897834873 hasConceptScore W2897834873C8642999 @default.
- W2897834873 hasLocation W28978348731 @default.
- W2897834873 hasOpenAccess W2897834873 @default.
- W2897834873 hasPrimaryLocation W28978348731 @default.
- W2897834873 hasRelatedWork W1949807611 @default.
- W2897834873 hasRelatedWork W2525236913 @default.
- W2897834873 hasRelatedWork W2593807705 @default.
- W2897834873 hasRelatedWork W2795133445 @default.
- W2897834873 hasRelatedWork W2803253185 @default.
- W2897834873 hasRelatedWork W2899462473 @default.
- W2897834873 hasRelatedWork W2947620140 @default.
- W2897834873 hasRelatedWork W2952183644 @default.
- W2897834873 hasRelatedWork W2962981549 @default.
- W2897834873 hasRelatedWork W2963916311 @default.
- W2897834873 hasRelatedWork W2964767742 @default.
- W2897834873 hasRelatedWork W2968558963 @default.
- W2897834873 hasRelatedWork W2990844796 @default.
- W2897834873 hasRelatedWork W2994881943 @default.
- W2897834873 hasRelatedWork W3102476541 @default.
- W2897834873 hasRelatedWork W3104340084 @default.
- W2897834873 hasRelatedWork W3120740533 @default.
- W2897834873 hasRelatedWork W3139019868 @default.
- W2897834873 hasRelatedWork W3180563047 @default.
- W2897834873 hasRelatedWork W3208413417 @default.
- W2897834873 isParatext "false" @default.
- W2897834873 isRetracted "false" @default.
- W2897834873 magId "2897834873" @default.
- W2897834873 workType "article" @default.