Matches in SemOpenAlex for { <https://semopenalex.org/work/W4319862430> ?p ?o ?g. }
Showing items 1 to 99 of
99
with 100 items per page.
- W4319862430 abstract "Deep learning Text-to-Speech (TTS) systems have achieved impressive generated speech quality, close to human parity. However, they suffer from training stability issues and in-correct alignment between the intermediate acoustic representation and the text input. In this work, we propose Regotron, a regularized Tacotron2 version which alleviates the training issues by augmenting the objective function with an additional term, which penalizes non-monotonic alignments in the location-sensitive attention mechanism. By introducing this regularization term we demonstrate its effectiveness to stabilize the training process, produce a monotonic attention quicker (13% of the total number of epochs compared to Tacotron2) and reduce the alignment errors during inference. Moreover, Regotron has minimal additional computational overhead, reduces common TTS mistakes and at the same time achieves improved speech naturalness according to subjective mean opinion scores (MOS) collected from 50 evaluators." @default.
- W4319862430 created "2023-02-11" @default.
- W4319862430 creator A5026032406 @default.
- W4319862430 creator A5035535505 @default.
- W4319862430 creator A5060812182 @default.
- W4319862430 creator A5061662829 @default.
- W4319862430 creator A5084949286 @default.
- W4319862430 creator A5087680388 @default.
- W4319862430 date "2023-01-09" @default.
- W4319862430 modified "2023-09-29" @default.
- W4319862430 title "Regotron: Regularizing the Tacotron2 Architecture Via Monotonic Alignment Loss" @default.
- W4319862430 cites W2767052532 @default.
- W4319862430 cites W2963300588 @default.
- W4319862430 cites W2963609956 @default.
- W4319862430 cites W2964243274 @default.
- W4319862430 cites W3015282541 @default.
- W4319862430 cites W3015478688 @default.
- W4319862430 cites W3015922793 @default.
- W4319862430 cites W3016136182 @default.
- W4319862430 cites W3095828473 @default.
- W4319862430 cites W3171066367 @default.
- W4319862430 cites W3195077602 @default.
- W4319862430 cites W3197294703 @default.
- W4319862430 cites W3198048682 @default.
- W4319862430 cites W3198633333 @default.
- W4319862430 cites W3198843334 @default.
- W4319862430 doi "https://doi.org/10.1109/slt54892.2023.10023268" @default.
- W4319862430 hasPublicationYear "2023" @default.
- W4319862430 type Work @default.
- W4319862430 citedByCount "0" @default.
- W4319862430 crossrefType "proceedings-article" @default.
- W4319862430 hasAuthorship W4319862430A5026032406 @default.
- W4319862430 hasAuthorship W4319862430A5035535505 @default.
- W4319862430 hasAuthorship W4319862430A5060812182 @default.
- W4319862430 hasAuthorship W4319862430A5061662829 @default.
- W4319862430 hasAuthorship W4319862430A5084949286 @default.
- W4319862430 hasAuthorship W4319862430A5087680388 @default.
- W4319862430 hasConcept C111919701 @default.
- W4319862430 hasConcept C119857082 @default.
- W4319862430 hasConcept C121332964 @default.
- W4319862430 hasConcept C134306372 @default.
- W4319862430 hasConcept C134537474 @default.
- W4319862430 hasConcept C154945302 @default.
- W4319862430 hasConcept C162324750 @default.
- W4319862430 hasConcept C176217482 @default.
- W4319862430 hasConcept C17744445 @default.
- W4319862430 hasConcept C199539241 @default.
- W4319862430 hasConcept C21547014 @default.
- W4319862430 hasConcept C2776135515 @default.
- W4319862430 hasConcept C2776214188 @default.
- W4319862430 hasConcept C2776359362 @default.
- W4319862430 hasConcept C2779960059 @default.
- W4319862430 hasConcept C28490314 @default.
- W4319862430 hasConcept C33923547 @default.
- W4319862430 hasConcept C41008148 @default.
- W4319862430 hasConcept C62520636 @default.
- W4319862430 hasConcept C62897895 @default.
- W4319862430 hasConcept C72169020 @default.
- W4319862430 hasConcept C94625758 @default.
- W4319862430 hasConceptScore W4319862430C111919701 @default.
- W4319862430 hasConceptScore W4319862430C119857082 @default.
- W4319862430 hasConceptScore W4319862430C121332964 @default.
- W4319862430 hasConceptScore W4319862430C134306372 @default.
- W4319862430 hasConceptScore W4319862430C134537474 @default.
- W4319862430 hasConceptScore W4319862430C154945302 @default.
- W4319862430 hasConceptScore W4319862430C162324750 @default.
- W4319862430 hasConceptScore W4319862430C176217482 @default.
- W4319862430 hasConceptScore W4319862430C17744445 @default.
- W4319862430 hasConceptScore W4319862430C199539241 @default.
- W4319862430 hasConceptScore W4319862430C21547014 @default.
- W4319862430 hasConceptScore W4319862430C2776135515 @default.
- W4319862430 hasConceptScore W4319862430C2776214188 @default.
- W4319862430 hasConceptScore W4319862430C2776359362 @default.
- W4319862430 hasConceptScore W4319862430C2779960059 @default.
- W4319862430 hasConceptScore W4319862430C28490314 @default.
- W4319862430 hasConceptScore W4319862430C33923547 @default.
- W4319862430 hasConceptScore W4319862430C41008148 @default.
- W4319862430 hasConceptScore W4319862430C62520636 @default.
- W4319862430 hasConceptScore W4319862430C62897895 @default.
- W4319862430 hasConceptScore W4319862430C72169020 @default.
- W4319862430 hasConceptScore W4319862430C94625758 @default.
- W4319862430 hasFunder F4320320300 @default.
- W4319862430 hasFunder F4320335322 @default.
- W4319862430 hasLocation W43198624301 @default.
- W4319862430 hasOpenAccess W4319862430 @default.
- W4319862430 hasPrimaryLocation W43198624301 @default.
- W4319862430 hasRelatedWork W2067665617 @default.
- W4319862430 hasRelatedWork W2109084804 @default.
- W4319862430 hasRelatedWork W2286917416 @default.
- W4319862430 hasRelatedWork W2511279186 @default.
- W4319862430 hasRelatedWork W2963879839 @default.
- W4319862430 hasRelatedWork W3194822862 @default.
- W4319862430 hasRelatedWork W4225130343 @default.
- W4319862430 hasRelatedWork W4297798732 @default.
- W4319862430 hasRelatedWork W4298157600 @default.
- W4319862430 hasRelatedWork W4319862430 @default.
- W4319862430 isParatext "false" @default.
- W4319862430 isRetracted "false" @default.
- W4319862430 workType "article" @default.