1 |
END-TO-END SPEECH RECOGNITION FROM FEDERATED ACOUSTIC MODELS
|
|
|
|
In: The International Conference on Acoustics, Speech, & Signal Processing (ICASSP) ; https://hal.archives-ouvertes.fr/hal-03601224 ; The International Conference on Acoustics, Speech, & Signal Processing (ICASSP), May 2022, Singapour, Singapore (2022)
|
|
Abstract:
International audience ; Training Automatic Speech Recognition (ASR) models under federated learning (FL) settings has attracted a lot of attention recently. However, the FL scenarios often presented in the literature are artificial and fail to capture the complexity of real FL systems. In this paper, we construct a challenging and realistic ASR federated experimental setup consisting of clients with heterogeneous data distributions using the French and Italian sets of the CommonVoice dataset, a large heterogeneous dataset containing thousands of different speakers, acoustic environments and noises. We present the first empirical study on attention-based sequence-to-sequence Endto-End (E2E) ASR model with three aggregation weighting strategies-standard FedAvg, loss-based aggregation and a novel word error rate (WER)-based aggregation, compared in two realistic FL scenarios: cross-silo with 10 clients and cross-device with 2K and 4K clients. Our analysis on E2E ASR from heterogeneous and realistic federated acoustic models provides the foundations for future research and development of realistic FL-based ASR applications.
|
|
Keyword:
[INFO]Computer Science [cs]
|
|
URL: https://hal.archives-ouvertes.fr/hal-03601224 https://hal.archives-ouvertes.fr/hal-03601224/file/2104.14297.pdf https://hal.archives-ouvertes.fr/hal-03601224/document
|
|
BASE
|
|
Hide details
|
|
2 |
Translating Headers of Tabular Data: A Pilot Study of Schema Translation ...
|
|
|
|
BASE
|
|
Show details
|
|
3 |
A Hybrid Semantic Parsing Approach for Tabular Data Analysis ...
|
|
|
|
BASE
|
|
Show details
|
|
5 |
Mobile technology utilization among patients from diverse cultural and linguistic backgrounds attending cardiac rehabilitation.
|
|
|
|
BASE
|
|
Show details
|
|
6 |
Textual organisation and construal of interpersonal meanings in different genres of medical texts ...
|
|
|
|
BASE
|
|
Show details
|
|
7 |
Textual organisation and construal of interpersonal meanings in different genres of medical texts
|
|
|
|
BASE
|
|
Show details
|
|
8 |
Mobile Technology Utilization Among Patients From Diverse Cultural and Linguistic Backgrounds Attending Cardiac Rehabilitation in Australia: Descriptive, Case-Matched Comparative Study
|
|
|
|
BASE
|
|
Show details
|
|
9 |
Why did you withdraw? Experiences of Chinese international doctoral students in Canada
|
|
|
|
BASE
|
|
Show details
|
|
10 |
Mobile Technology Use Across Age Groups in Patients Eligible for Cardiac Rehabilitation: Survey Study
|
|
|
|
BASE
|
|
Show details
|
|
11 |
A Tentative Research on Chinese Culture Integrated Into College English Teaching: Taking an English Optional Course Dialogue With Chinese Culture as an Example
|
|
|
|
In: Cross-Cultural Communication; Vol 12, No 11 (2016): Cross-Cultural Communication; 27-32 ; 1923-6700 ; 1712-8358 (2016)
|
|
BASE
|
|
Show details
|
|
|
|