BYOL-S: Learning Self-supervised Speech Representations by Bootstrapping
Research output: Contribution to conference › Paper › Research
Standard
BYOL-S : Learning Self-supervised Speech Representations by Bootstrapping. / Elbanna, Gasser; Scheidwasser-Clow, Neil; Kegler, Mikolaj; Beckmann, Pierre; Hajal, Karl El; Cernak, Milos.
2022. Paper presented at HEAR: Holistic Evaluation of Audio Representations (NeurIPS 2021 Competition).Research output: Contribution to conference › Paper › Research
Harvard
APA
Vancouver
Author
Bibtex
}
RIS
TY - CONF
T1 - BYOL-S
AU - Elbanna, Gasser
AU - Scheidwasser-Clow, Neil
AU - Kegler, Mikolaj
AU - Beckmann, Pierre
AU - Hajal, Karl El
AU - Cernak, Milos
PY - 2022/6/24
Y1 - 2022/6/24
N2 - Methods for extracting audio and speech features have been studied since pioneering work on spectrum analysis decades ago. Recent efforts are guided by the ambition to develop general-purpose audio representations. For example, deep neural networks can extract optimal embeddings if they are trained on large audio datasets. This work extends existing methods based on self-supervised learning by bootstrapping, proposes various encoder architectures, and explores the effects of using different pre-training datasets. Lastly, we present a novel training framework to come up with a hybrid audio representation, which combines handcrafted and data-driven learned audio features. All the proposed representations were evaluated within the HEAR NeurIPS 2021 challenge for auditory scene classification and timestamp detection tasks. Our results indicate that the hybrid model with a convolutional transformer as the encoder yields superior performance in most HEAR challenge tasks.
AB - Methods for extracting audio and speech features have been studied since pioneering work on spectrum analysis decades ago. Recent efforts are guided by the ambition to develop general-purpose audio representations. For example, deep neural networks can extract optimal embeddings if they are trained on large audio datasets. This work extends existing methods based on self-supervised learning by bootstrapping, proposes various encoder architectures, and explores the effects of using different pre-training datasets. Lastly, we present a novel training framework to come up with a hybrid audio representation, which combines handcrafted and data-driven learned audio features. All the proposed representations were evaluated within the HEAR NeurIPS 2021 challenge for auditory scene classification and timestamp detection tasks. Our results indicate that the hybrid model with a convolutional transformer as the encoder yields superior performance in most HEAR challenge tasks.
KW - cs.SD
KW - cs.AI
KW - cs.LG
KW - eess.AS
M3 - Paper
Y2 - 13 December 2021 through 14 December 2021
ER -
ID: 337591641