Matches in SemOpenAlex for { <https://semopenalex.org/work/W4287331382> ?p ?o ?g. }
Showing items 1 to 66 of
66
with 100 items per page.
- W4287331382 abstract "Performance degradation of an Automatic Speech Recognition (ASR) system is commonly observed when the test acoustic condition is different from training. Hence, it is essential to make ASR systems robust against various environmental distortions, such as background noises and reverberations. In a multi-stream paradigm, improving robustness takes account of handling a variety of unseen single-stream conditions and inter-stream dynamics. Previously, a practical two-stage training strategy was proposed within multi-stream end-to-end ASR, where Stage-2 formulates the multi-stream model with features from Stage-1 Universal Feature Extractor (UFE). In this paper, as an extension, we introduce a two-stage augmentation scheme focusing on mismatch scenarios: Stage-1 Augmentation aims to address single-stream input varieties with data augmentation techniques; Stage-2 Time Masking applies temporal masks on UFE features of randomly selected streams to simulate diverse stream combinations. During inference, we also present adaptive Connectionist Temporal Classification (CTC) fusion with the help of hierarchical attention mechanisms. Experiments have been conducted on two datasets, DIRHA and AMI, as a multi-stream scenario. Compared with the previous training strategy, substantial improvements are reported with relative word error rate reductions of 29.7-59.3% across several unseen stream combinations." @default.
- W4287331382 created "2022-07-25" @default.
- W4287331382 creator A5042260050 @default.
- W4287331382 creator A5046596198 @default.
- W4287331382 creator A5079096032 @default.
- W4287331382 date "2021-02-05" @default.
- W4287331382 modified "2023-10-16" @default.
- W4287331382 title "Two-Stage Augmentation and Adaptive CTC Fusion for Improved Robustness of Multi-Stream End-to-End ASR" @default.
- W4287331382 hasPublicationYear "2021" @default.
- W4287331382 type Work @default.
- W4287331382 citedByCount "0" @default.
- W4287331382 crossrefType "posted-content" @default.
- W4287331382 hasAuthorship W4287331382A5042260050 @default.
- W4287331382 hasAuthorship W4287331382A5046596198 @default.
- W4287331382 hasAuthorship W4287331382A5079096032 @default.
- W4287331382 hasBestOaLocation W42873313821 @default.
- W4287331382 hasConcept C104317684 @default.
- W4287331382 hasConcept C117978034 @default.
- W4287331382 hasConcept C127413603 @default.
- W4287331382 hasConcept C153180895 @default.
- W4287331382 hasConcept C154945302 @default.
- W4287331382 hasConcept C185592680 @default.
- W4287331382 hasConcept C21880701 @default.
- W4287331382 hasConcept C2776214188 @default.
- W4287331382 hasConcept C2778484313 @default.
- W4287331382 hasConcept C28490314 @default.
- W4287331382 hasConcept C41008148 @default.
- W4287331382 hasConcept C50644808 @default.
- W4287331382 hasConcept C55493867 @default.
- W4287331382 hasConcept C63479239 @default.
- W4287331382 hasConcept C74296488 @default.
- W4287331382 hasConcept C76155785 @default.
- W4287331382 hasConcept C8521452 @default.
- W4287331382 hasConceptScore W4287331382C104317684 @default.
- W4287331382 hasConceptScore W4287331382C117978034 @default.
- W4287331382 hasConceptScore W4287331382C127413603 @default.
- W4287331382 hasConceptScore W4287331382C153180895 @default.
- W4287331382 hasConceptScore W4287331382C154945302 @default.
- W4287331382 hasConceptScore W4287331382C185592680 @default.
- W4287331382 hasConceptScore W4287331382C21880701 @default.
- W4287331382 hasConceptScore W4287331382C2776214188 @default.
- W4287331382 hasConceptScore W4287331382C2778484313 @default.
- W4287331382 hasConceptScore W4287331382C28490314 @default.
- W4287331382 hasConceptScore W4287331382C41008148 @default.
- W4287331382 hasConceptScore W4287331382C50644808 @default.
- W4287331382 hasConceptScore W4287331382C55493867 @default.
- W4287331382 hasConceptScore W4287331382C63479239 @default.
- W4287331382 hasConceptScore W4287331382C74296488 @default.
- W4287331382 hasConceptScore W4287331382C76155785 @default.
- W4287331382 hasConceptScore W4287331382C8521452 @default.
- W4287331382 hasLocation W42873313821 @default.
- W4287331382 hasOpenAccess W4287331382 @default.
- W4287331382 hasPrimaryLocation W42873313821 @default.
- W4287331382 hasRelatedWork W13920906 @default.
- W4287331382 hasRelatedWork W1403418 @default.
- W4287331382 hasRelatedWork W14254504 @default.
- W4287331382 hasRelatedWork W1919069 @default.
- W4287331382 hasRelatedWork W2279739 @default.
- W4287331382 hasRelatedWork W3703470 @default.
- W4287331382 hasRelatedWork W6882942 @default.
- W4287331382 hasRelatedWork W8315361 @default.
- W4287331382 hasRelatedWork W8459508 @default.
- W4287331382 hasRelatedWork W9420449 @default.
- W4287331382 isParatext "false" @default.
- W4287331382 isRetracted "false" @default.
- W4287331382 workType "article" @default.