Title:
ELECTRONIC DEVICE AND METHOD OF LOW LATENCY SPEECH ENHANCEMENT USING AUTOREGRESSIVE CONDITIONING-BASED NEURAL NETWORK MODEL
Document Type and Number:
WIPO Patent Application WO/2024/080699
Kind Code:
A1
Abstract:
A neural method model is trained by, in an initial training iteration, training the neural network model in a teacher forcing mode in which an autoregressive channel includes a ground-truth shifted waveform, and outputting predictions of the neural network model; and in at least one additional training iteration, replacing the ground-truth shifted waveform in the autoregressive channel with the predictions of the neural network model obtained in a previous training iteration. An inference may then be performed by providing, for the neural network model, an additional channel containing at least one prediction of the neural network model outputted during training; and performing speech enhancement using the neural network model.
More Like This:
Inventors:
BABAEV NIKOLAS ANDREW (RU)
ANDREEV PAVEL KONSTANTINOVICH (RU)
SAGINBAEV AZAT RUSTAMOVICH (RU)
SHCHEKOTOV IVAN SERGEEVICH (RU)
ANDREEV PAVEL KONSTANTINOVICH (RU)
SAGINBAEV AZAT RUSTAMOVICH (RU)
SHCHEKOTOV IVAN SERGEEVICH (RU)
Application Number:
PCT/KR2023/015526
Publication Date:
April 18, 2024
Filing Date:
October 10, 2023
Export Citation:
Assignee:
SAMSUNG ELECTRONICS CO LTD (KR)
International Classes:
G10L21/02; G06N3/02; G10L15/16
Foreign References:
| CN112634174A | 2021-04-09 | |||
| US20220309651A1 | 2022-09-29 |
Other References:
HEHE FAN: "PointRNN: Point Recurrent Neural Network for Moving Point Cloud Processing", ARXIV.ORG, CORNELL UNIVERSITY LIBRARY, ARXIV.ORG, ITHACA, 24 November 2019 (2019-11-24), Ithaca, XP093159503, Retrieved from the Internet DOI: 10.48550/arxiv.1910.08287
JONATHAN SHEN; RUOMING PANG; RON J. WEISS; MIKE SCHUSTER; NAVDEEP JAITLY; ZONGHENG YANG; ZHIFENG CHEN; YU ZHANG; YUXUAN WANG: "Natural tts synthesis by conditioning wavenet on mel spectrogram predictions", 2018 IEEE INTERNATIONAL CONFERENCE ON ACOUSTICS, SPEECH AND SIGNAL PROCESSING (ICASSP), IEEE, 1 January 2018 (2018-01-01), pages 1 - 5, XP002806894, DOI: 10.1109/ICASSP.2018.8461368
YIJIN LIU: "Confidence-Aware Scheduled Sampling for Neural Machine Translation", FINDINGS OF THE ASSOCIATION FOR COMPUTATIONAL LINGUISTICS: ACL-IJCNLP 2021, ASSOCIATION FOR COMPUTATIONAL LINGUISTICS, STROUDSBURG, PA, USA, 1 January 2021 (2021-01-01), Stroudsburg, PA, USA, pages 2327 - 2337, XP093159505, DOI: 10.18653/v1/2021.findings-acl.205
JONATHAN SHEN; RUOMING PANG; RON J. WEISS; MIKE SCHUSTER; NAVDEEP JAITLY; ZONGHENG YANG; ZHIFENG CHEN; YU ZHANG; YUXUAN WANG: "Natural tts synthesis by conditioning wavenet on mel spectrogram predictions", 2018 IEEE INTERNATIONAL CONFERENCE ON ACOUSTICS, SPEECH AND SIGNAL PROCESSING (ICASSP), IEEE, 1 January 2018 (2018-01-01), pages 1 - 5, XP002806894, DOI: 10.1109/ICASSP.2018.8461368
YIJIN LIU: "Confidence-Aware Scheduled Sampling for Neural Machine Translation", FINDINGS OF THE ASSOCIATION FOR COMPUTATIONAL LINGUISTICS: ACL-IJCNLP 2021, ASSOCIATION FOR COMPUTATIONAL LINGUISTICS, STROUDSBURG, PA, USA, 1 January 2021 (2021-01-01), Stroudsburg, PA, USA, pages 2327 - 2337, XP093159505, DOI: 10.18653/v1/2021.findings-acl.205
Attorney, Agent or Firm:
KIM, Tae-hun et al. (KR)
Download PDF:
Previous Patent: LIPID NANOPARTICLE COMPOSITION COMPRISING SODIUM DNA
Next Patent: METHOD AND APPARATUS FOR CONFIGURING AN UL CODEBOOK
Next Patent: METHOD AND APPARATUS FOR CONFIGURING AN UL CODEBOOK