中科微点-全球专利检索

  1. Login
  2. Sign Up

Search Results

13 records (took 0.014 seconds)

    Separating speech by source in audio recordings by predicting isolated audio signals conditioned on speaker representations
    1.
    发明授权
    Separating speech by source in audio recordings by predicting isolated audio signals conditioned on speaker representations 有权

    公开(公告)号：US12236970B2

    公开(公告)日：2025-02-25

    申请号：US17967726

    申请日：2022-10-17

    Applicant: Google LLC

    Inventor： Neil Zeghidour ,  David Grangier

    IPC: G10L21/028 ,  G06N3/045 ,  G06N3/08 ,  G10L17/04 ,  G10L17/18 ,  G10L21/0208 ,  G10L21/0272 ,  G10L21/0316 ,  G10L25/30

    Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for performing speech separation. One of the methods includes obtaining a recording comprising speech from a plurality of speakers; processing the recording using a speaker neural network having speaker parameter values and configured to process the recording in accordance with the speaker parameter values to generate a plurality of per-recording speaker representations, each speaker representation representing features of a respective identified speaker in the recording; and processing the per-recording speaker representations and the recording using a separation neural network having separation parameter values and configured to process the recording and the speaker representations in accordance with the separation parameter values to generate, for each speaker representation, a respective predicted isolated audio signal that corresponds to speech of one of the speakers in the recording.

    END-TO-END SPEECH DIARIZATION VIA ITERATIVE SPEAKER EMBEDDING
    2.
    发明公开
    END-TO-END SPEECH DIARIZATION VIA ITERATIVE SPEAKER EMBEDDING 审中-公开

    公开(公告)号：US20240144957A1

    公开(公告)日：2024-05-02

    申请号：US18544647

    申请日：2023-12-19

    Applicant: Google LLC

    Inventor： David Grangier ,  Neil Zeghidour ,  Oliver Teboul

    IPC: G10L25/78 ,  G06N3/04 ,  G10L15/06 ,  G10L15/07 ,  G10L17/18 ,  G10L19/008

    CPC classification number: G10L25/78 ,  G06N3/04 ,  G10L15/063 ,  G10L15/07 ,  G10L17/18 ,  G10L19/008

    Abstract: A method includes receiving an input audio signal corresponding to utterances spoken by multiple speakers. The method also includes encoding the input audio signal into a sequence of T temporal embeddings. During each of a plurality of iterations each corresponding to a respective speaker of the multiple speakers, the method includes selecting a respective speaker embedding for the respective speaker by determining a probability that the corresponding temporal embedding includes a presence of voice activity by a single new speaker for which a speaker embedding was not previously selected during a previous iteration and selecting the respective speaker embedding for the respective speaker as the temporal embedding. The method also includes, at each time step, predicting a respective voice activity indicator for each respective speaker of the multiple speakers based on the respective speaker embeddings selected during the plurality of iterations and the temporal embedding.

    Minimum Bayes Risk Decoding with Neural Quality Metrics
    4.
    发明公开
    Minimum Bayes Risk Decoding with Neural Quality Metrics 审中-公开

    公开(公告)号：US20230259759A1

    公开(公告)日：2023-08-17

    申请号：US17673714

    申请日：2022-02-16

    Applicant: Google LLC

    Inventor： Qijun Tan ,  Markus Freitag ,  David Grangier

    IPC: G06N3/08

    CPC classification number: G06N3/08

    Abstract: Provided are systems and methods for sequence-to-sequence modeling with neural quality metrics. More particularly, example aspects of the present disclosure relate to minimum bayes risk (MBR) decoding with neural metrics for machine translation. According to example aspects of the present disclosure, a set of candidate outputs can be sampled from a machine translation model given a source sequence. Given the set of candidate outputs, systems and methods according to example aspects of the present disclosure can select a hypothesis with high expected utility with respect to the distribution over a set of pseudo-references from the machine translation model.

    End-To-End Speech Diarization Via Iterative Speaker Embedding
    5.
    发明申请
    End-To-End Speech Diarization Via Iterative Speaker Embedding 有权

    公开(公告)号：US20220375492A1

    公开(公告)日：2022-11-24

    申请号：US17304514

    申请日：2021-06-22

    Applicant: Google LLC

    Inventor： David Grangier ,  Neil Zeghidour ,  Oliver Teboul

    IPC: G10L25/78 ,  G10L19/008 ,  G06N3/04 ,  G10L15/07 ,  G10L15/06 ,  G10L17/18

    Abstract: A method includes receiving an input audio signal corresponding to utterances spoken by multiple speakers. The method also includes encoding the input audio signal into a sequence of T temporal embeddings. During each of a plurality of iterations each corresponding to a respective speaker of the multiple speakers, the method includes selecting a respective speaker embedding for the respective speaker by determining a probability that the corresponding temporal embedding includes a presence of voice activity by a single new speaker for which a speaker embedding was not previously selected during a previous iteration and selecting the respective speaker embedding for the respective speaker as the temporal embedding. The method also includes, at each time step, predicting a respective voice activity indicator for each respective speaker of the multiple speakers based on the respective speaker embeddings selected during the plurality of iterations and the temporal embedding.

    Separating speech by source in audio recordings by predicting isolated audio signals conditioned on speaker representations
    8.
    发明授权
    Separating speech by source in audio recordings by predicting isolated audio signals conditioned on speaker representations 有权

    公开(公告)号：US11475909B2

    公开(公告)日：2022-10-18

    申请号：US17170657

    申请日：2021-02-08

    Applicant: Google LLC

    Inventor： Neil Zeghidour ,  David Grangier

    IPC: G10L21/028 ,  G10L21/0316 ,  G10L17/04 ,  G10L17/18 ,  G06N3/04 ,  G06N3/08

    Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for performing speech separation. One of the methods includes obtaining a recording comprising speech from a plurality of speakers; processing the recording using a speaker neural network having speaker parameter values and configured to process the recording in accordance with the speaker parameter values to generate a plurality of per-recording speaker representations, each speaker representation representing features of a respective identified speaker in the recording; and processing the per-recording speaker representations and the recording using a separation neural network having separation parameter values and configured to process the recording and the speaker representations in accordance with the separation parameter values to generate, for each speaker representation, a respective predicted isolated audio signal that corresponds to speech of one of the speakers in the recording.

    Learning Strides in Convolutional Neural Networks
    9.
    发明申请
    Learning Strides in Convolutional Neural Networks 有权

    公开(公告)号：US20250005354A1

    公开(公告)日：2025-01-02

    申请号：US18698691

    申请日：2022-10-05

    Applicant: Google LLC

    Inventor： Neil Zeghidour ,  Rachid Riad ,  Olivier Teboul ,  David Grangier

    IPC: G06N3/08

    Abstract: A method of training a machine learning model, includes receiving training data for the machine learning model, wherein the training data comprises a plurality of batches. The method also includes applying a downsampling layer of the machine learning model to the plurality of batches of the training data to determine a stride comprising a learnable parameter for the downsampling layer. Applying the downsampling layer of the machine learning model to a batch of the training data includes projecting an input in a spatial domain to a Fourier domain, constructing a mask in the Fourier domain based on a current value of the stride and dimensions of the input, applying the mask as a low-pass filter to the projected input to produce a tensor in the Fourier domain, cropping the tensor based on the mask, and transforming the cropped tensor to the spatial domain.

    End-to-end speech diarization via iterative speaker embedding
    10.
    发明授权
    End-to-end speech diarization via iterative speaker embedding 有权

    公开(公告)号：US11887623B2

    公开(公告)日：2024-01-30

    申请号：US17304514

    申请日：2021-06-22

    Applicant: Google LLC

    Inventor： David Grangier ,  Neil Zeghidour ,  Oliver Teboul

    IPC: G10L25/78 ,  G06N3/04 ,  G10L15/06 ,  G10L15/07 ,  G10L17/18 ,  G10L19/008

    CPC classification number: G10L25/78 ,  G06N3/04 ,  G10L15/063 ,  G10L15/07 ,  G10L17/18 ,  G10L19/008

    Abstract: A method includes receiving an input audio signal corresponding to utterances spoken by multiple speakers. The method also includes encoding the input audio signal into a sequence of T temporal embeddings. During each of a plurality of iterations each corresponding to a respective speaker of the multiple speakers, the method includes selecting a respective speaker embedding for the respective speaker by determining a probability that the corresponding temporal embedding includes a presence of voice activity by a single new speaker for which a speaker embedding was not previously selected during a previous iteration and selecting the respective speaker embedding for the respective speaker as the temporal embedding. The method also includes, at each time step, predicting a respective voice activity indicator for each respective speaker of the multiple speakers based on the respective speaker embeddings selected during the plurality of iterations and the temporal embedding.

Patent Agency Ranking