Matches in SemOpenAlex for { <https://semopenalex.org/work/W3171516653> ?p ?o ?g. }
Showing items 1 to 85 of
85
with 100 items per page.
- W3171516653 endingPage "1104" @default.
- W3171516653 startingPage "1092" @default.
- W3171516653 abstract "Responding to user data deletion requests, removing noisy examples, or deleting corrupted training data are just a few reasons for wanting to delete instances from a machine learning (ML) model. However, efficiently removing this data from an ML model is generally difficult. In this paper, we introduce data removal-enabled (DaRE) forests, a variant of random forests that enables the removal of training data with minimal retraining. Model updates for each DaRE tree in the forest are exact, meaning that removing instances from a DaRE model yields exactly the same model as retraining from scratch on updated data. DaRE trees use randomness and caching to make data deletion efficient. The upper levels of DaRE trees use random nodes, which choose split attributes and thresholds uniformly at random. These nodes rarely require updates because they only minimally depend on the data. At the lower levels, splits are chosen to greedily optimize a split criterion such as Gini index or mutual information. DaRE trees cache statistics at each node and training data at each leaf, so that only the necessary subtrees are updated as data is removed. For numerical attributes, greedy nodes optimize over a random subset of thresholds, so that they can maintain statistics while approximating the optimal threshold. By adjusting the number of thresholds considered for greedy nodes, and the number of random nodes, DaRE trees can trade off between more accurate predictions and more efficient updates. In experiments on 13 real-world datasets and one synthetic dataset, we find DaRE forests delete data orders of magnitude faster than retraining from scratch while sacrificing little to no predictive power." @default.
- W3171516653 created "2021-06-22" @default.
- W3171516653 creator A5053373401 @default.
- W3171516653 creator A5086131462 @default.
- W3171516653 date "2021-07-18" @default.
- W3171516653 modified "2023-09-26" @default.
- W3171516653 title "Machine Unlearning for Random Forests" @default.
- W3171516653 hasPublicationYear "2021" @default.
- W3171516653 type Work @default.
- W3171516653 sameAs 3171516653 @default.
- W3171516653 citedByCount "5" @default.
- W3171516653 countsByYear W31715166532021 @default.
- W3171516653 countsByYear W31715166532023 @default.
- W3171516653 crossrefType "proceedings-article" @default.
- W3171516653 hasAuthorship W3171516653A5053373401 @default.
- W3171516653 hasAuthorship W3171516653A5086131462 @default.
- W3171516653 hasConcept C101722063 @default.
- W3171516653 hasConcept C105795698 @default.
- W3171516653 hasConcept C111919701 @default.
- W3171516653 hasConcept C113174947 @default.
- W3171516653 hasConcept C11413529 @default.
- W3171516653 hasConcept C115537543 @default.
- W3171516653 hasConcept C119857082 @default.
- W3171516653 hasConcept C124101348 @default.
- W3171516653 hasConcept C125112378 @default.
- W3171516653 hasConcept C127413603 @default.
- W3171516653 hasConcept C134306372 @default.
- W3171516653 hasConcept C144133560 @default.
- W3171516653 hasConcept C154945302 @default.
- W3171516653 hasConcept C155202549 @default.
- W3171516653 hasConcept C169258074 @default.
- W3171516653 hasConcept C2778712577 @default.
- W3171516653 hasConcept C33923547 @default.
- W3171516653 hasConcept C41008148 @default.
- W3171516653 hasConcept C62611344 @default.
- W3171516653 hasConcept C66938386 @default.
- W3171516653 hasConceptScore W3171516653C101722063 @default.
- W3171516653 hasConceptScore W3171516653C105795698 @default.
- W3171516653 hasConceptScore W3171516653C111919701 @default.
- W3171516653 hasConceptScore W3171516653C113174947 @default.
- W3171516653 hasConceptScore W3171516653C11413529 @default.
- W3171516653 hasConceptScore W3171516653C115537543 @default.
- W3171516653 hasConceptScore W3171516653C119857082 @default.
- W3171516653 hasConceptScore W3171516653C124101348 @default.
- W3171516653 hasConceptScore W3171516653C125112378 @default.
- W3171516653 hasConceptScore W3171516653C127413603 @default.
- W3171516653 hasConceptScore W3171516653C134306372 @default.
- W3171516653 hasConceptScore W3171516653C144133560 @default.
- W3171516653 hasConceptScore W3171516653C154945302 @default.
- W3171516653 hasConceptScore W3171516653C155202549 @default.
- W3171516653 hasConceptScore W3171516653C169258074 @default.
- W3171516653 hasConceptScore W3171516653C2778712577 @default.
- W3171516653 hasConceptScore W3171516653C33923547 @default.
- W3171516653 hasConceptScore W3171516653C41008148 @default.
- W3171516653 hasConceptScore W3171516653C62611344 @default.
- W3171516653 hasConceptScore W3171516653C66938386 @default.
- W3171516653 hasLocation W31715166531 @default.
- W3171516653 hasOpenAccess W3171516653 @default.
- W3171516653 hasPrimaryLocation W31715166531 @default.
- W3171516653 hasRelatedWork W1488996941 @default.
- W3171516653 hasRelatedWork W1543525280 @default.
- W3171516653 hasRelatedWork W1836929938 @default.
- W3171516653 hasRelatedWork W1959785497 @default.
- W3171516653 hasRelatedWork W1989218322 @default.
- W3171516653 hasRelatedWork W2006131099 @default.
- W3171516653 hasRelatedWork W2012180059 @default.
- W3171516653 hasRelatedWork W2023618245 @default.
- W3171516653 hasRelatedWork W2031217601 @default.
- W3171516653 hasRelatedWork W2083256575 @default.
- W3171516653 hasRelatedWork W2088519218 @default.
- W3171516653 hasRelatedWork W2122671957 @default.
- W3171516653 hasRelatedWork W2143959471 @default.
- W3171516653 hasRelatedWork W2335881461 @default.
- W3171516653 hasRelatedWork W2985381883 @default.
- W3171516653 hasRelatedWork W3013112461 @default.
- W3171516653 hasRelatedWork W3035644192 @default.
- W3171516653 hasRelatedWork W3173117233 @default.
- W3171516653 hasRelatedWork W3194064817 @default.
- W3171516653 hasRelatedWork W40476960 @default.
- W3171516653 isParatext "false" @default.
- W3171516653 isRetracted "false" @default.
- W3171516653 magId "3171516653" @default.
- W3171516653 workType "article" @default.