Zobrazeno 1 - 10
of 3 249
pro vyhledávání: '"Kingsbury, P."'
Autor:
Cui, Xiaodong, Saif, A F M, Lu, Songtao, Chen, Lisha, Chen, Tianyi, Kingsbury, Brian, Saon, George
In this paper, we propose a bilevel joint unsupervised and supervised training (BL-JUST) framework for automatic speech recognition. Compared to the conventional pre-training and fine-tuning strategy which is a disconnected two-stage process, BL-JUST
Externí odkaz:
http://arxiv.org/abs/2412.08548
Background: Hip replacement procedures improve patient lives by relieving pain and restoring mobility. Predicting hip replacement in advance could reduce pain by enabling timely interventions, prioritising individuals for surgery or rehabilitation, a
Externí odkaz:
http://arxiv.org/abs/2409.06585
Autor:
Kingsbury, Nathaniel
Let $R$ be a finite ring and define the paraboloid $P = \{(x_1, \dots, x_d)\in R^d|x_d = x_1^2 + \dots + x_{d-1}^2\}.$ Suppose that for a sequence of finite rings of size tending to infinity, the Fourier transform of $P$ satisfies a square-root law f
Externí odkaz:
http://arxiv.org/abs/2405.13248
The emergence of industrial-scale speech recognition (ASR) models such as Whisper and USM, trained on 1M hours of weakly labelled and 12M hours of audio only proprietary data respectively, has led to a stronger need for large scale public ASR corpora
Externí odkaz:
http://arxiv.org/abs/2402.00235
Joint Unsupervised and Supervised Training for Automatic Speech Recognition via Bilevel Optimization
In this paper, we present a novel bilevel optimization-based training approach to training acoustic models for automatic speech recognition (ASR) tasks that we term {bi-level joint unsupervised and supervised training (BL-JUST)}. {BL-JUST employs a l
Externí odkaz:
http://arxiv.org/abs/2401.06980
Autor:
Das, Sanskriti, Rickel, Mary, Leroy, Adam, Pingel, Nickolas M., Pisano, D. J., Heald, George, Mathur, Smita, Kingsbury, Joshua, Sardone, Amy
We probe the neutral circumgalactic medium (CGM) along the major axes of NGC891 and NGC4565 in 21-cm emission out to $\gtrsim 100$kpc using the Green Bank Telescope (GBT), extending our previous minor axes observations. We achieve an unprecedented $5
Externí odkaz:
http://arxiv.org/abs/2312.06880
Soft random sampling (SRS) is a simple yet effective approach for efficient training of large-scale deep neural networks when dealing with massive data. SRS selects a subset uniformly at random with replacement from the full data set in each epoch. I
Externí odkaz:
http://arxiv.org/abs/2311.12727
Autor:
Gulotta, Alessandro, Polimeni, Marco, Lenton, Samuel, Starr, Charles G., Kingsbury, Jonathan S., Stradner, Anna, Zaccarelli, Emanuela, Schurtenberger, Peter
Charges and their contribution to protein-protein interactions are essential for the key structural and dynamic properties of monoclonal antibody (mAb) solutions. In fact, they influence the apparent molecular weight, the static structure factor, the
Externí odkaz:
http://arxiv.org/abs/2311.01986
Autor:
Anna M. Anderson, Lucy Brading, Laura Swaithes, Nicola Evans, Sophia E. Fedorowicz, Darren Murinas, Elizabeth Atkinson, Alice Moult, Tatum Yip, Parveen Ayub, Krysia Dziedzic, Philip G. Conaghan, Gretl A. McHugh, Amy Rebane, Sarah R. Kingsbury
Publikováno v:
Research Involvement and Engagement, Vol 10, Iss 1, Pp 1-22 (2024)
Abstract Background Certain groups are commonly under-served by health research due to exclusionary models of research design/delivery. Working in partnership with under-served groups is key to improving inclusion. This project aimed to explore the u
Externí odkaz:
https://doaj.org/article/da103cb2186a4dada4b79bba1626343c
Non-autoregressive (NAR) modeling has gained significant interest in speech processing since these models achieve dramatically lower inference time than autoregressive (AR) models while also achieving good transcription accuracy. Since NAR automatic
Externí odkaz:
http://arxiv.org/abs/2309.10926