FRINK Linked Data Fragments server

Matches in SemOpenAlex for { <https://semopenalex.org/work/W3093603503> ?p ?o ?g. }

• W3093603503 abstract "This paper studies the problem of approximately unlearning a Bayesian model from a small subset of the training data to be erased. We frame this problem as one of minimizing the Kullback-Leibler divergence between the approximate posterior belief of model parameters after directly unlearning from erased data vs. the exact posterior belief from retraining with remaining data. Using the variational inference (VI) framework, we show that it is equivalent to minimizing an evidence upper bound which trades off between fully unlearning from erased data vs. not entirely forgetting the posterior belief given the full data (i.e., including the remaining data); the latter prevents catastrophic unlearning that can render the model useless. In model training with VI, only an approximate (instead of exact) posterior belief given the full data can be obtained, which makes unlearning even more challenging. We propose two novel tricks to tackle this challenge. We empirically demonstrate our unlearning methods on Bayesian models such as sparse Gaussian process and logistic regression using synthetic and real-world datasets." @default.
• W3093603503 created "2020-10-29" @default.
• W3093603503 creator A5003654479 @default.
• W3093603503 creator A5018942019 @default.
• W3093603503 creator A5030304400 @default.
• W3093603503 date "2020-10-24" @default.
• W3093603503 modified "2023-09-27" @default.
• W3093603503 title "Variational Bayesian Unlearning." @default.
• W3093603503 cites W1488996941 @default.
• W3093603503 cites W1571870753 @default.
• W3093603503 cites W1663973292 @default.
• W3093603503 cites W1746819321 @default.
• W3093603503 cites W1777124189 @default.
• W3093603503 cites W1865714576 @default.
• W3093603503 cites W1886178467 @default.
• W3093603503 cites W1952209694 @default.
• W3093603503 cites W2255848560 @default.
• W3093603503 cites W2474046462 @default.
• W3093603503 cites W2587284713 @default.
• W3093603503 cites W2945038936 @default.
• W3093603503 cites W2962974546 @default.
• W3093603503 cites W2963047245 @default.
• W3093603503 cites W2963055046 @default.
• W3093603503 cites W2963090522 @default.
• W3093603503 cites W2963479359 @default.
• W3093603503 cites W2963487581 @default.
• W3093603503 cites W2963542673 @default.
• W3093603503 cites W2964153655 @default.
• W3093603503 cites W2964206586 @default.
• W3093603503 cites W2964276430 @default.
• W3093603503 cites W2970695928 @default.
• W3093603503 cites W2971124187 @default.
• W3093603503 cites W2977291151 @default.
• W3093603503 cites W2985671575 @default.
• W3093603503 cites W2988337058 @default.
• W3093603503 cites W2997782896 @default.
• W3093603503 cites W3035644192 @default.
• W3093603503 cites W3124806946 @default.
• W3093603503 hasPublicationYear "2020" @default.
• W3093603503 type Work @default.
• W3093603503 sameAs 3093603503 @default.
• W3093603503 citedByCount "1" @default.
• W3093603503 countsByYear W30936035032021 @default.
• W3093603503 crossrefType "posted-content" @default.
• W3093603503 hasAuthorship W3093603503A5003654479 @default.
• W3093603503 hasAuthorship W3093603503A5018942019 @default.
• W3093603503 hasAuthorship W3093603503A5030304400 @default.
• W3093603503 hasConcept C107673813 @default.
• W3093603503 hasConcept C119857082 @default.
• W3093603503 hasConcept C121332964 @default.
• W3093603503 hasConcept C138885662 @default.
• W3093603503 hasConcept C154945302 @default.
• W3093603503 hasConcept C15744967 @default.
• W3093603503 hasConcept C160234255 @default.
• W3093603503 hasConcept C163716315 @default.
• W3093603503 hasConcept C180747234 @default.
• W3093603503 hasConcept C207390915 @default.
• W3093603503 hasConcept C2776214188 @default.
• W3093603503 hasConcept C41008148 @default.
• W3093603503 hasConcept C41895202 @default.
• W3093603503 hasConcept C57830394 @default.
• W3093603503 hasConcept C61326573 @default.
• W3093603503 hasConcept C62520636 @default.
• W3093603503 hasConcept C7149132 @default.
• W3093603503 hasConceptScore W3093603503C107673813 @default.
• W3093603503 hasConceptScore W3093603503C119857082 @default.
• W3093603503 hasConceptScore W3093603503C121332964 @default.
• W3093603503 hasConceptScore W3093603503C138885662 @default.
• W3093603503 hasConceptScore W3093603503C154945302 @default.
• W3093603503 hasConceptScore W3093603503C15744967 @default.
• W3093603503 hasConceptScore W3093603503C160234255 @default.
• W3093603503 hasConceptScore W3093603503C163716315 @default.
• W3093603503 hasConceptScore W3093603503C180747234 @default.
• W3093603503 hasConceptScore W3093603503C207390915 @default.
• W3093603503 hasConceptScore W3093603503C2776214188 @default.
• W3093603503 hasConceptScore W3093603503C41008148 @default.
• W3093603503 hasConceptScore W3093603503C41895202 @default.
• W3093603503 hasConceptScore W3093603503C57830394 @default.
• W3093603503 hasConceptScore W3093603503C61326573 @default.
• W3093603503 hasConceptScore W3093603503C62520636 @default.
• W3093603503 hasConceptScore W3093603503C7149132 @default.
• W3093603503 hasLocation W30936035031 @default.
• W3093603503 hasOpenAccess W3093603503 @default.
• W3093603503 hasPrimaryLocation W30936035031 @default.
• W3093603503 hasRelatedWork W1486664518 @default.
• W3093603503 hasRelatedWork W1493199654 @default.
• W3093603503 hasRelatedWork W1815331012 @default.
• W3093603503 hasRelatedWork W2013327003 @default.
• W3093603503 hasRelatedWork W2128547631 @default.
• W3093603503 hasRelatedWork W2171349590 @default.
• W3093603503 hasRelatedWork W224960998 @default.
• W3093603503 hasRelatedWork W2254392641 @default.
• W3093603503 hasRelatedWork W2739647525 @default.
• W3093603503 hasRelatedWork W2747180671 @default.
• W3093603503 hasRelatedWork W2785953858 @default.
• W3093603503 hasRelatedWork W2807557496 @default.
• W3093603503 hasRelatedWork W2916197674 @default.
• W3093603503 hasRelatedWork W2947490203 @default.
• W3093603503 hasRelatedWork W2949701357 @default.
• W3093603503 hasRelatedWork W3022749775 @default.

• next