FRINK Linked Data Fragments server

Matches in SemOpenAlex for { <https://semopenalex.org/work/W2950304420> ?p ?o ?g. }

• W2950304420 abstract "Recurrent Neural Networks can be trained to produce sequences of tokens given some input, as exemplified by recent results in machine translation and image captioning. The current approach to training them consists of maximizing the likelihood of each token in the sequence given the current (recurrent) state and the previous token. At inference, the unknown previous token is then replaced by a token generated by the model itself. This discrepancy between training and inference can yield errors that can accumulate quickly along the generated sequence. We propose a curriculum learning strategy to gently change the training process from a fully guided scheme using the true previous token, towards a less guided scheme which mostly uses the generated token instead. Experiments on several sequence prediction tasks show that this approach yields significant improvements. Moreover, it was used successfully in our winning entry to the MSCOCO image captioning challenge, 2015." @default.
• W2950304420 created "2019-06-27" @default.
• W2950304420 creator A5003562101 @default.
• W2950304420 creator A5017529415 @default.
• W2950304420 creator A5021878400 @default.
• W2950304420 creator A5031967550 @default.
• W2950304420 date "2015-06-09" @default.
• W2950304420 modified "2023-09-27" @default.
• W2950304420 title "Scheduled Sampling for Sequence Prediction with Recurrent Neural Networks" @default.
• W2950304420 cites W1524333225 @default.
• W2950304420 cites W1527575280 @default.
• W2950304420 cites W1586532344 @default.
• W2950304420 cites W1861492603 @default.
• W2950304420 cites W1869752048 @default.
• W2950304420 cites W1895577753 @default.
• W2950304420 cites W1905882502 @default.
• W2950304420 cites W1931639407 @default.
• W2950304420 cites W1947481528 @default.
• W2950304420 cites W1956340063 @default.
• W2950304420 cites W2064675550 @default.
• W2950304420 cites W2088911157 @default.
• W2950304420 cites W2104917081 @default.
• W2950304420 cites W2147880316 @default.
• W2950304420 cites W2268617045 @default.
• W2950304420 cites W2296073425 @default.
• W2950304420 cites W2734523219 @default.
• W2950304420 cites W2949117887 @default.
• W2950304420 cites W2949888546 @default.
• W2950304420 cites W2952122856 @default.
• W2950304420 cites W2962706528 @default.
• W2950304420 cites W2962957031 @default.
• W2950304420 cites W2964308564 @default.
• W2950304420 cites W595069947 @default.
• W2950304420 hasPublicationYear "2015" @default.
• W2950304420 type Work @default.
• W2950304420 sameAs 2950304420 @default.
• W2950304420 citedByCount "238" @default.
• W2950304420 countsByYear W29503044202015 @default.
• W2950304420 countsByYear W29503044202016 @default.
• W2950304420 countsByYear W29503044202017 @default.
• W2950304420 countsByYear W29503044202018 @default.
• W2950304420 countsByYear W29503044202019 @default.
• W2950304420 countsByYear W29503044202020 @default.
• W2950304420 countsByYear W29503044202021 @default.
• W2950304420 crossrefType "posted-content" @default.
• W2950304420 hasAuthorship W2950304420A5003562101 @default.
• W2950304420 hasAuthorship W2950304420A5017529415 @default.
• W2950304420 hasAuthorship W2950304420A5021878400 @default.
• W2950304420 hasAuthorship W2950304420A5031967550 @default.
• W2950304420 hasConcept C108583219 @default.
• W2950304420 hasConcept C111919701 @default.
• W2950304420 hasConcept C115067241 @default.
• W2950304420 hasConcept C115961682 @default.
• W2950304420 hasConcept C119857082 @default.
• W2950304420 hasConcept C134306372 @default.
• W2950304420 hasConcept C147168706 @default.
• W2950304420 hasConcept C154945302 @default.
• W2950304420 hasConcept C157657479 @default.
• W2950304420 hasConcept C2776214188 @default.
• W2950304420 hasConcept C2778112365 @default.
• W2950304420 hasConcept C28490314 @default.
• W2950304420 hasConcept C31258907 @default.
• W2950304420 hasConcept C33923547 @default.
• W2950304420 hasConcept C41008148 @default.
• W2950304420 hasConcept C48145219 @default.
• W2950304420 hasConcept C50644808 @default.
• W2950304420 hasConcept C54355233 @default.
• W2950304420 hasConcept C77618280 @default.
• W2950304420 hasConcept C86803240 @default.
• W2950304420 hasConcept C98045186 @default.
• W2950304420 hasConceptScore W2950304420C108583219 @default.
• W2950304420 hasConceptScore W2950304420C111919701 @default.
• W2950304420 hasConceptScore W2950304420C115067241 @default.
• W2950304420 hasConceptScore W2950304420C115961682 @default.
• W2950304420 hasConceptScore W2950304420C119857082 @default.
• W2950304420 hasConceptScore W2950304420C134306372 @default.
• W2950304420 hasConceptScore W2950304420C147168706 @default.
• W2950304420 hasConceptScore W2950304420C154945302 @default.
• W2950304420 hasConceptScore W2950304420C157657479 @default.
• W2950304420 hasConceptScore W2950304420C2776214188 @default.
• W2950304420 hasConceptScore W2950304420C2778112365 @default.
• W2950304420 hasConceptScore W2950304420C28490314 @default.
• W2950304420 hasConceptScore W2950304420C31258907 @default.
• W2950304420 hasConceptScore W2950304420C33923547 @default.
• W2950304420 hasConceptScore W2950304420C41008148 @default.
• W2950304420 hasConceptScore W2950304420C48145219 @default.
• W2950304420 hasConceptScore W2950304420C50644808 @default.
• W2950304420 hasConceptScore W2950304420C54355233 @default.
• W2950304420 hasConceptScore W2950304420C77618280 @default.
• W2950304420 hasConceptScore W2950304420C86803240 @default.
• W2950304420 hasConceptScore W2950304420C98045186 @default.
• W2950304420 hasLocation W29503044201 @default.
• W2950304420 hasOpenAccess W2950304420 @default.
• W2950304420 hasPrimaryLocation W29503044201 @default.
• W2950304420 hasRelatedWork W1895577753 @default.
• W2950304420 hasRelatedWork W1905882502 @default.
• W2950304420 hasRelatedWork W1924770834 @default.
• W2950304420 hasRelatedWork W1956340063 @default.
• W2950304420 hasRelatedWork W2016589492 @default.
• W2950304420 hasRelatedWork W2064675550 @default.

• next