Matches in SemOpenAlex for { <https://semopenalex.org/work/W3172007248> ?p ?o ?g. }
Showing items 1 to 91 of
91
with 100 items per page.
- W3172007248 endingPage "4595" @default.
- W3172007248 startingPage "4585" @default.
- W3172007248 abstract "Direct loss minimization is a popular approach for learning predictors over structured label spaces. This approach is computationally appealing as it replaces integration with optimization and allows to propagate gradients in a deep net using loss-perturbed prediction. Recently, this technique was extended to generative models, while introducing a randomized predictor that samples a structure from a randomly perturbed score function. In this work, we learn the variance of these randomized structured predictors and show that it balances better between the learned score function and the randomized noise in structured prediction. We demonstrate empirically the effectiveness of learning the balance between the signal and the random noise in structured discrete spaces." @default.
- W3172007248 created "2021-06-22" @default.
- W3172007248 creator A5000106487 @default.
- W3172007248 creator A5016682466 @default.
- W3172007248 date "2021-07-18" @default.
- W3172007248 modified "2023-09-23" @default.
- W3172007248 title "Learning Randomly Perturbed Structured Predictors for Direct Loss Minimization" @default.
- W3172007248 hasPublicationYear "2021" @default.
- W3172007248 type Work @default.
- W3172007248 sameAs 3172007248 @default.
- W3172007248 citedByCount "0" @default.
- W3172007248 crossrefType "proceedings-article" @default.
- W3172007248 hasAuthorship W3172007248A5000106487 @default.
- W3172007248 hasAuthorship W3172007248A5016682466 @default.
- W3172007248 hasConcept C105795698 @default.
- W3172007248 hasConcept C107321475 @default.
- W3172007248 hasConcept C11413529 @default.
- W3172007248 hasConcept C115961682 @default.
- W3172007248 hasConcept C119857082 @default.
- W3172007248 hasConcept C121955636 @default.
- W3172007248 hasConcept C14036430 @default.
- W3172007248 hasConcept C141071460 @default.
- W3172007248 hasConcept C144133560 @default.
- W3172007248 hasConcept C147764199 @default.
- W3172007248 hasConcept C154945302 @default.
- W3172007248 hasConcept C155108698 @default.
- W3172007248 hasConcept C167966045 @default.
- W3172007248 hasConcept C168563851 @default.
- W3172007248 hasConcept C196083921 @default.
- W3172007248 hasConcept C199360897 @default.
- W3172007248 hasConcept C22367795 @default.
- W3172007248 hasConcept C33923547 @default.
- W3172007248 hasConcept C39890363 @default.
- W3172007248 hasConcept C41008148 @default.
- W3172007248 hasConcept C65660741 @default.
- W3172007248 hasConcept C71924100 @default.
- W3172007248 hasConcept C78458016 @default.
- W3172007248 hasConcept C86803240 @default.
- W3172007248 hasConcept C99498987 @default.
- W3172007248 hasConceptScore W3172007248C105795698 @default.
- W3172007248 hasConceptScore W3172007248C107321475 @default.
- W3172007248 hasConceptScore W3172007248C11413529 @default.
- W3172007248 hasConceptScore W3172007248C115961682 @default.
- W3172007248 hasConceptScore W3172007248C119857082 @default.
- W3172007248 hasConceptScore W3172007248C121955636 @default.
- W3172007248 hasConceptScore W3172007248C14036430 @default.
- W3172007248 hasConceptScore W3172007248C141071460 @default.
- W3172007248 hasConceptScore W3172007248C144133560 @default.
- W3172007248 hasConceptScore W3172007248C147764199 @default.
- W3172007248 hasConceptScore W3172007248C154945302 @default.
- W3172007248 hasConceptScore W3172007248C155108698 @default.
- W3172007248 hasConceptScore W3172007248C167966045 @default.
- W3172007248 hasConceptScore W3172007248C168563851 @default.
- W3172007248 hasConceptScore W3172007248C196083921 @default.
- W3172007248 hasConceptScore W3172007248C199360897 @default.
- W3172007248 hasConceptScore W3172007248C22367795 @default.
- W3172007248 hasConceptScore W3172007248C33923547 @default.
- W3172007248 hasConceptScore W3172007248C39890363 @default.
- W3172007248 hasConceptScore W3172007248C41008148 @default.
- W3172007248 hasConceptScore W3172007248C65660741 @default.
- W3172007248 hasConceptScore W3172007248C71924100 @default.
- W3172007248 hasConceptScore W3172007248C78458016 @default.
- W3172007248 hasConceptScore W3172007248C86803240 @default.
- W3172007248 hasConceptScore W3172007248C99498987 @default.
- W3172007248 hasOpenAccess W3172007248 @default.
- W3172007248 hasRelatedWork W1497745584 @default.
- W3172007248 hasRelatedWork W2253831701 @default.
- W3172007248 hasRelatedWork W2405231250 @default.
- W3172007248 hasRelatedWork W2410344874 @default.
- W3172007248 hasRelatedWork W2516383060 @default.
- W3172007248 hasRelatedWork W2604702452 @default.
- W3172007248 hasRelatedWork W2617020555 @default.
- W3172007248 hasRelatedWork W2747927423 @default.
- W3172007248 hasRelatedWork W2781406523 @default.
- W3172007248 hasRelatedWork W2782586160 @default.
- W3172007248 hasRelatedWork W2895600735 @default.
- W3172007248 hasRelatedWork W2996774258 @default.
- W3172007248 hasRelatedWork W3007361404 @default.
- W3172007248 hasRelatedWork W3015298730 @default.
- W3172007248 hasRelatedWork W3037206323 @default.
- W3172007248 hasRelatedWork W3042060288 @default.
- W3172007248 hasRelatedWork W3083953438 @default.
- W3172007248 hasRelatedWork W3109605628 @default.
- W3172007248 hasRelatedWork W3133628563 @default.
- W3172007248 hasRelatedWork W3139111359 @default.
- W3172007248 isParatext "false" @default.
- W3172007248 isRetracted "false" @default.
- W3172007248 magId "3172007248" @default.
- W3172007248 workType "article" @default.