FRINK Linked Data Fragments server

Matches in SemOpenAlex for { <https://semopenalex.org/work/W4281617617> ?p ?o ?g. }

• W4281617617 abstract "Embeddings, low-dimensional vector representation of objects, are fundamental in building modern machine learning systems. In industrial settings, there is usually an embedding team that trains an embedding model to solve intended tasks (e.g., product recommendation). The produced embeddings are then widely consumed by consumer teams to solve their unintended tasks (e.g., fraud detection). However, as the embedding model gets updated and retrained to improve performance on the intended task, the newly-generated embeddings are no longer compatible with the existing consumer models. This means that historical versions of the embeddings can never be retired or all consumer teams have to retrain their models to make them compatible with the latest version of the embeddings, both of which are extremely costly in practice. Here we study the problem of embedding version updates and their backward compatibility. We formalize the problem where the goal is for the embedding team to keep updating the embedding version, while the consumer teams do not have to retrain their models. We develop a solution based on learning backward compatible embeddings, which allows the embedding model version to be updated frequently, while also allowing the latest version of the embedding to be quickly transformed into any backward compatible historical version of it, so that consumer teams do not have to retrain their models. Under our framework, we explore six methods and systematically evaluate them on a real-world recommender system application. We show that the best method, which we call BC-Aligner, maintains backward compatibility with existing unintended tasks even after multiple model version updates. Simultaneously, BC-Aligner achieves the intended task performance similar to the embedding model that is solely optimized for the intended task." @default.
• W4281617617 created "2022-06-13" @default.
• W4281617617 creator A5021259488 @default.
• W4281617617 creator A5022336015 @default.
• W4281617617 creator A5057079625 @default.
• W4281617617 creator A5081428282 @default.
• W4281617617 creator A5091272738 @default.
• W4281617617 creator A5091515272 @default.
• W4281617617 date "2022-08-14" @default.
• W4281617617 modified "2023-10-10" @default.
• W4281617617 title "Learning Backward Compatible Embeddings" @default.
• W4281617617 cites W2294774419 @default.
• W4281617617 cites W2561995736 @default.
• W4281617617 cites W2962756421 @default.
• W4281617617 cites W2971196067 @default.
• W4281617617 cites W3034594226 @default.
• W4281617617 cites W3100278010 @default.
• W4281617617 cites W3101767658 @default.
• W4281617617 cites W3104723404 @default.
• W4281617617 cites W3171903345 @default.
• W4281617617 cites W4214521899 @default.
• W4281617617 doi "https://doi.org/10.1145/3534678.3539194" @default.
• W4281617617 hasPublicationYear "2022" @default.
• W4281617617 type Work @default.
• W4281617617 citedByCount "6" @default.
• W4281617617 countsByYear W42816176172023 @default.
• W4281617617 crossrefType "proceedings-article" @default.
• W4281617617 hasAuthorship W4281617617A5021259488 @default.
• W4281617617 hasAuthorship W4281617617A5022336015 @default.
• W4281617617 hasAuthorship W4281617617A5057079625 @default.
• W4281617617 hasAuthorship W4281617617A5081428282 @default.
• W4281617617 hasAuthorship W4281617617A5091272738 @default.
• W4281617617 hasAuthorship W4281617617A5091515272 @default.
• W4281617617 hasBestOaLocation W42816176172 @default.
• W4281617617 hasConcept C119857082 @default.
• W4281617617 hasConcept C127413603 @default.
• W4281617617 hasConcept C154945302 @default.
• W4281617617 hasConcept C201995342 @default.
• W4281617617 hasConcept C2780451532 @default.
• W4281617617 hasConcept C41008148 @default.
• W4281617617 hasConcept C41608201 @default.
• W4281617617 hasConcept C557471498 @default.
• W4281617617 hasConcept C80444323 @default.
• W4281617617 hasConceptScore W4281617617C119857082 @default.
• W4281617617 hasConceptScore W4281617617C127413603 @default.
• W4281617617 hasConceptScore W4281617617C154945302 @default.
• W4281617617 hasConceptScore W4281617617C201995342 @default.
• W4281617617 hasConceptScore W4281617617C2780451532 @default.
• W4281617617 hasConceptScore W4281617617C41008148 @default.
• W4281617617 hasConceptScore W4281617617C41608201 @default.
• W4281617617 hasConceptScore W4281617617C557471498 @default.
• W4281617617 hasConceptScore W4281617617C80444323 @default.
• W4281617617 hasLocation W42816176171 @default.
• W4281617617 hasLocation W42816176172 @default.
• W4281617617 hasOpenAccess W4281617617 @default.
• W4281617617 hasPrimaryLocation W42816176171 @default.
• W4281617617 hasRelatedWork W2045871438 @default.
• W4281617617 hasRelatedWork W2122731942 @default.
• W4281617617 hasRelatedWork W2348159088 @default.
• W4281617617 hasRelatedWork W2350747448 @default.
• W4281617617 hasRelatedWork W2368095327 @default.
• W4281617617 hasRelatedWork W2402445420 @default.
• W4281617617 hasRelatedWork W2499363748 @default.
• W4281617617 hasRelatedWork W2809363009 @default.
• W4281617617 hasRelatedWork W2968745142 @default.
• W4281617617 hasRelatedWork W2977378428 @default.
• W4281617617 isParatext "false" @default.
• W4281617617 isRetracted "false" @default.
• W4281617617 workType "article" @default.