Zobrazeno 1 - 10 of 91 pro vyhledávání: '"Jasha Droppo"'
1
Guided Contrastive Self-Supervised Pre-Training for Automatic Speech Recognition
Autor: Aparna Khare, Minhua Wu, Saurabhchand Bhati, Jasha Droppo, Roland Maas
Publikováno v: 2022 IEEE Spoken Language Technology Workshop (SLT).
Contrastive Predictive Coding (CPC) is a representation learning method that maximizes the mutual information between intermediate latent representations and the output of a given model. It can be used to effectively initialize the encoder of an Auto
Externí odkaz: https://explore.openaire.eu/search/publication?articleId=doi_dedup___::9d87dba96d5c6b4e9997d37f3e7c61e5
https://doi.org/10.1109/slt54892.2023.10022676
Zobrazit plný text záznamu
3
Improving fairness in speaker verification via Group-adapted Fusion Network
Autor: Hua Shen, Yuguang Yang, Guoli Sun, Ryan Langman, Eunjung Han, Jasha Droppo, Andreas Stolcke
Modern speaker verification models use deep neural networks to encode utterance audio into discriminative embedding vectors. During the training process, these networks are typically optimized to differentiate arbitrary speakers. This learning proces
Externí odkaz: https://explore.openaire.eu/search/publication?articleId=doi_dedup___::3c65d26d8b49c5d1b0b0b12399e458c5
Zobrazit plný text záznamu
4
Scaling Laws for Acoustic Models
Autor: Oguz H. Elibol, Jasha Droppo
Publikováno v: Interspeech 2021.
There is a recent trend in machine learning to increase model quality by growing models to sizes previously thought to be unreasonable. Recent work has shown that autoregressive generative models with cross-entropy objective functions exhibit smooth
Externí odkaz: https://explore.openaire.eu/search/publication?articleId=doi_________::99e60dcafdd99c301a68a9a749b57ce7
https://doi.org/10.21437/interspeech.2021-1644
Zobrazit plný text záznamu
5
CoDERT: Distilling Encoder Representations with Co-Learning for Transducer-Based Speech Recognition
Autor: Rupak Vignesh Swaminathan, Athanasios Mouchtaris, Brian King, Jasha Droppo, Grant P. Strimel
Publikováno v: Interspeech 2021.
We propose a simple yet effective method to compress an RNN-Transducer (RNN-T) through the well-known knowledge distillation paradigm. We show that the transducer's encoder outputs naturally have a high entropy and contain rich information about acou
Externí odkaz: https://explore.openaire.eu/search/publication?articleId=doi_dedup___::0c1c4bad7b8047c783a853aabf28d6cb
https://doi.org/10.21437/interspeech.2021-797
Zobrazit plný text záznamu
7
Improving Multi-Speaker TTS Prosody Variance with a Residual Encoder and Normalizing Flows
Autor: Roberto Barra-Chicote, Jasha Droppo, Grzegorz Beringer, Julian Roth, Iván Vallés-Pérez
Publikováno v: Interspeech 2021.
Text-to-speech systems recently achieved almost indistinguishable quality from human speech. However, the prosody of those systems is generally flatter than natural speech, producing samples with low expressiveness. Disentanglement of speaker id and
Externí odkaz: https://explore.openaire.eu/search/publication?articleId=doi_dedup___::e1d19fd06dcf5573b5303fcf03cfca9e
https://doi.org/10.21437/interspeech.2021-562
Zobrazit plný text záznamu
9
SynthASR: Unlocking Synthetic Data for Speech Recognition
Autor: Roland Maas, Jasha Droppo, Roberto Barra-Chicote, Yixiong Meng, Amin Fazel, Wei Yang, Yulan Liu
End-to-end (E2E) automatic speech recognition (ASR) models have recently demonstrated superior performance over the traditional hybrid ASR models. Training an E2E ASR model requires a large amount of data which is not only expensive but may also rais
Externí odkaz: https://explore.openaire.eu/search/publication?articleId=doi_dedup___::7c7ab09241c6ea402622a7b43a7286d8
http://arxiv.org/abs/2106.07803
Zobrazit plný text záznamu
10
Exploring the application of synthetic audio in training keyword spotters
Autor: Jasha Droppo, Peng Liu, Yuriy Mishchenko, Anish Shah, Roberto Barra Chicote, Jeff Condal, Andrew Werchniak
Publikováno v: ICASSP
The study of keyword spotting, a subfield within the broader field of speech recognition that centers around identifying individual keywords in speech audio, has gained particular importance in recent years with the rise of personal voice assistants
Externí odkaz: https://explore.openaire.eu/search/publication?articleId=doi_________::974c9d4c23fd6740fdba3e4dc0529591
https://doi.org/10.1109/icassp39728.2021.9413448
Zobrazit plný text záznamu

Vyhledávací nástroje:

Upřesnit hledání

Omezení vyhledávání
Plný text Recenzováno Digitální knihovna AV ČR
Zdroje