Matches in SemOpenAlex for { <https://semopenalex.org/work/W4386875889> ?p ?o ?g. }
Showing items 1 to 79 of
79
with 100 items per page.
- W4386875889 abstract "Diffusion models have shown promising results in speech enhancement, using a task-adapted diffusion process for the conditional generation of clean speech given a noisy mixture. However, at test time, the neural network used for score estimation is called multiple times to solve the iterative reverse process. This results in a slow inference process and causes discretization errors that accumulate over the sampling trajectory. In this paper, we address these limitations through a two-stage training approach. In the first stage, we train the diffusion model the usual way using the generative denoising score matching loss. In the second stage, we compute the enhanced signal by solving the reverse process and compare the resulting estimate to the clean speech target using a predictive loss. We show that using this second training stage enables achieving the same performance as the baseline model using only 5 function evaluations instead of 60 function evaluations. While the performance of usual generative diffusion algorithms drops dramatically when lowering the number of function evaluations (NFEs) to obtain single-step diffusion, we show that our proposed method keeps a steady performance and therefore largely outperforms the diffusion baseline in this setting and also generalizes better than its predictive counterpart." @default.
- W4386875889 created "2023-09-20" @default.
- W4386875889 creator A5006784595 @default.
- W4386875889 creator A5027458841 @default.
- W4386875889 creator A5087017732 @default.
- W4386875889 creator A5087022569 @default.
- W4386875889 date "2023-09-18" @default.
- W4386875889 modified "2023-09-27" @default.
- W4386875889 title "Single and Few-step Diffusion for Generative Speech Enhancement" @default.
- W4386875889 doi "https://doi.org/10.48550/arxiv.2309.09677" @default.
- W4386875889 hasPublicationYear "2023" @default.
- W4386875889 type Work @default.
- W4386875889 citedByCount "0" @default.
- W4386875889 crossrefType "posted-content" @default.
- W4386875889 hasAuthorship W4386875889A5006784595 @default.
- W4386875889 hasAuthorship W4386875889A5027458841 @default.
- W4386875889 hasAuthorship W4386875889A5087017732 @default.
- W4386875889 hasAuthorship W4386875889A5087022569 @default.
- W4386875889 hasBestOaLocation W43868758891 @default.
- W4386875889 hasConcept C105795698 @default.
- W4386875889 hasConcept C111919701 @default.
- W4386875889 hasConcept C11413529 @default.
- W4386875889 hasConcept C121332964 @default.
- W4386875889 hasConcept C134306372 @default.
- W4386875889 hasConcept C14036430 @default.
- W4386875889 hasConcept C154945302 @default.
- W4386875889 hasConcept C163294075 @default.
- W4386875889 hasConcept C165064840 @default.
- W4386875889 hasConcept C2776182073 @default.
- W4386875889 hasConcept C2776214188 @default.
- W4386875889 hasConcept C3017618536 @default.
- W4386875889 hasConcept C33923547 @default.
- W4386875889 hasConcept C41008148 @default.
- W4386875889 hasConcept C56739046 @default.
- W4386875889 hasConcept C68710425 @default.
- W4386875889 hasConcept C69357855 @default.
- W4386875889 hasConcept C73000952 @default.
- W4386875889 hasConcept C78458016 @default.
- W4386875889 hasConcept C86803240 @default.
- W4386875889 hasConcept C97355855 @default.
- W4386875889 hasConcept C98045186 @default.
- W4386875889 hasConceptScore W4386875889C105795698 @default.
- W4386875889 hasConceptScore W4386875889C111919701 @default.
- W4386875889 hasConceptScore W4386875889C11413529 @default.
- W4386875889 hasConceptScore W4386875889C121332964 @default.
- W4386875889 hasConceptScore W4386875889C134306372 @default.
- W4386875889 hasConceptScore W4386875889C14036430 @default.
- W4386875889 hasConceptScore W4386875889C154945302 @default.
- W4386875889 hasConceptScore W4386875889C163294075 @default.
- W4386875889 hasConceptScore W4386875889C165064840 @default.
- W4386875889 hasConceptScore W4386875889C2776182073 @default.
- W4386875889 hasConceptScore W4386875889C2776214188 @default.
- W4386875889 hasConceptScore W4386875889C3017618536 @default.
- W4386875889 hasConceptScore W4386875889C33923547 @default.
- W4386875889 hasConceptScore W4386875889C41008148 @default.
- W4386875889 hasConceptScore W4386875889C56739046 @default.
- W4386875889 hasConceptScore W4386875889C68710425 @default.
- W4386875889 hasConceptScore W4386875889C69357855 @default.
- W4386875889 hasConceptScore W4386875889C73000952 @default.
- W4386875889 hasConceptScore W4386875889C78458016 @default.
- W4386875889 hasConceptScore W4386875889C86803240 @default.
- W4386875889 hasConceptScore W4386875889C97355855 @default.
- W4386875889 hasConceptScore W4386875889C98045186 @default.
- W4386875889 hasLocation W43868758891 @default.
- W4386875889 hasOpenAccess W4386875889 @default.
- W4386875889 hasPrimaryLocation W43868758891 @default.
- W4386875889 hasRelatedWork W146891625 @default.
- W4386875889 hasRelatedWork W2029297756 @default.
- W4386875889 hasRelatedWork W2277210191 @default.
- W4386875889 hasRelatedWork W2360394982 @default.
- W4386875889 hasRelatedWork W2365297781 @default.
- W4386875889 hasRelatedWork W2373515517 @default.
- W4386875889 hasRelatedWork W2386767533 @default.
- W4386875889 hasRelatedWork W2386796262 @default.
- W4386875889 hasRelatedWork W4205962296 @default.
- W4386875889 hasRelatedWork W2012842278 @default.
- W4386875889 isParatext "false" @default.
- W4386875889 isRetracted "false" @default.
- W4386875889 workType "article" @default.