公开(公告)号：US20250022477A1

公开(公告)日：2025-01-16

申请号：US18278746

申请日：2023-03-16

Applicant: Google LLC

Inventor： Ahmed Omran ,  Neil Zeghidour ,  Zalán Borsos ,  Félix de Chaumont Quitry ,  Marco Tagliasacchi

IPC: G10L19/038 ,  G10L25/30 ,  G10L25/60

Abstract: Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for training an encoder neural network and a decoder neural network. In one aspect, a method includes obtaining a first initial audio waveform and a first noisy audio waveform, obtaining a second initial audio waveform and a second noisy audio waveform, processing the first noisy audio waveform and the second noisy audio waveform using an encoder neural network, generating a blended embedding by concatenating: (i) clean feature dimensions from an embedding of the first noisy audio waveform, and (ii) noise feature dimensions from an embedding of the second noisy audio waveform, processing the blended embedding using a decoder neural network to generate a reconstructed audio waveform, determining gradients of an objective function; and updating parameter values of the encoder neural network and the decoder neural network using the gradients.

公开(公告)号：US20220383112A1

公开(公告)日：2022-12-01

申请号：US17764005

申请日：2020-09-23

Applicant: Google LLC

Inventor： Marco Tagliasacchi ,  Félix de Chaumont Quitry ,  Dominik Roblek

IPC: G06N3/08 ,  G06N3/04 ,  G10L25/30

Abstract: A system including a multi-task adapter neural network for performing multiple machine learning tasks is described. The adapter neural network is configured to receive a shared input for the machine learning tasks, and process the shared input to generate, for each of the machine learning tasks, a respective predicted output. The adapter neural network includes (i) a shared encoder configured to receive the shared input and to process the shared input to extract shared feature representations for the machine learning tasks, and (ii) multiple task-adapter encoders, each of the task-adapter encoders being associated with a respective machine learning task in the machine learning tasks and configured to: receive the shared input, receive the shared feature representations from the shared encoder, and process the shared input and the shared feature representations to generate the respective predicted output for the respective machine learning task.

公开(公告)号：US20210056980A1

公开(公告)日：2021-02-25

申请号：US16548146

申请日：2019-08-22

Applicant: Google LLC

Inventor： Beat Gfeller ,  Dominik Roblek ,  Félix de Chaumont Quitry ,  Marco Tagliasacchi

IPC: G10L19/035 ,  G10L25/18 ,  G10L19/038 ,  G06N20/00

Abstract: Systems and methods for training a machine-learned model are provided. A method can include can include obtaining an unlabeled audio signal, sampling the unlabeled audio signal to select one or more sampled slices, inputting the one or more sampled slices into a machine-learned model, receiving, as an output of the machine-learned model, one or more determined characteristics associated with the audio signal, determining a loss function for the machine-learned model based at least in part on a difference between the one or more determined characteristics and one or more corresponding ground truth characteristics of the audio signal, and training the machine-learned model from end to end based at least in part on the loss function. The one or more determined characteristics can include one or more reconstructed portions of the audio signal temporally adjacent to the one or more sampled slices or an estimated distance between two sampled slices.

公开(公告)号：US12190896B2

公开(公告)日：2025-01-07

申请号：US17856292

申请日：2022-07-01

Applicant: Google LLC

Inventor： Yunpeng Li ,  Marco Tagliasacchi ,  Dominik Roblek ,  Félix de Chaumont Quitry ,  Beat Gfeller ,  Hannah Raphaelle Muckenhirn ,  Victor Ungureanu ,  Oleg Rybakov ,  Karolis Misiunas ,  Zalán Borsos

IPC: G10L19/022 ,  G06N3/045

Abstract: Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for processing an input audio waveform using a generator neural network to generate an output audio waveform. In one aspect, a method comprises: receiving an input audio waveform; processing the input audio waveform using an encoder neural network to generate a set of feature vectors representing the input audio waveform; and processing the set of feature vectors representing the input audio waveform using a decoder neural network to generate an output audio waveform that comprises a respective output audio sample for each of a plurality of output time steps.

公开(公告)号：US20230377561A1

公开(公告)日：2023-11-23

申请号：US18029843

申请日：2021-10-04

Applicant: Google LLC

Inventor： Neil Zeghidour ,  Olivier Teboul ,  Félix de Chaumont Quitry ,  Marco Tagliasacchi

IPC: G10L15/02 ,  G10L15/16

CPC classification number: G10L15/02 ,  G10L15/16

Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for processing audio inputs using a learned audio frontend machine learning model that processes the audio input to generate a representation of the audio input. The representation can then be processed by an audio understanding model to generate a respective output for each of one or more audio understanding tasks.

公开(公告)号：US20230085596A1

公开(公告)日：2023-03-16

申请号：US17986477

申请日：2022-11-14

Applicant: Google LLC

Inventor： Beat Gfeller ,  Dominik Roblek ,  Félix de Chaumont Quitry ,  Marco Tagliasacchi

IPC: G10L19/035 ,  G06N20/00 ,  G10L19/038 ,  G10L25/18

Abstract: Systems and methods for training a machine-learned model are provided. A method can include can include obtaining an unlabeled audio signal, sampling the unlabeled audio signal to select one or more sampled slices, inputting the one or more sampled slices into a machine-learned model, receiving, as an output of the machine-learned model, one or more determined characteristics associated with the audio signal, determining a loss function for the machine-learned model based at least in part on a difference between the one or more determined characteristics and one or more corresponding ground truth characteristics of the audio signal, and training the machine-learned model from end to end based at least in part on the loss function. The one or more determined characteristics can include one or more reconstructed portions of the audio signal temporally adjacent to the one or more sampled slices or an estimated distance between two sampled slices.

公开(公告)号：US20230013370A1

公开(公告)日：2023-01-19

申请号：US17856292

申请日：2022-07-01

Applicant: Google LLC

Inventor： Yunpeng Li ,  Marco Tagliasacchi ,  Dominik Roblek ,  Félix de Chaumont Quitry ,  Beat Gfeller ,  Hannah Raphaelle Muckenhirn ,  Victor Ungureanu ,  Oleg Rybakov ,  Karolis Misiunas ,  Zalán Borsos

IPC: G10L19/022 ,  G06N3/04

Abstract: Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for processing an input audio waveform using a generator neural network to generate an output audio waveform. In one aspect, a method comprises: receiving an input audio waveform; processing the input audio waveform using an encoder neural network to generate a set of feature vectors representing the input audio waveform; and processing the set of feature vectors representing the input audio waveform using a decoder neural network to generate an output audio waveform that comprises a respective output audio sample for each of a plurality of output time steps.

公开(公告)号：US12165663B2

公开(公告)日：2024-12-10

申请号：US17986477

申请日：2022-11-14

Applicant: Google LLC

Inventor： Beat Gfeller ,  Dominik Roblek ,  Félix de Chaumont Quitry ,  Marco Tagliasacchi

IPC: G10L19/035 ,  G06N20/00 ,  G10L19/038 ,  G10L25/18

Abstract: Systems and methods for training a machine-learned model are provided. A method can include can include obtaining an unlabeled audio signal, sampling the unlabeled audio signal to select one or more sampled slices, inputting the one or more sampled slices into a machine-learned model, receiving, as an output of the machine-learned model, one or more determined characteristics associated with the audio signal, determining a loss function for the machine-learned model based at least in part on a difference between the one or more determined characteristics and one or more corresponding ground truth characteristics of the audio signal, and training the machine-learned model from end to end based at least in part on the loss function. The one or more determined characteristics can include one or more reconstructed portions of the audio signal temporally adjacent to the one or more sampled slices or an estimated distance between two sampled slices.

公开(公告)号：US11501787B2

公开(公告)日：2022-11-15

申请号：US16548146

申请日：2019-08-22

Applicant: Google LLC

Inventor： Beat Gfeller ,  Dominik Roblek ,  Félix de Chaumont Quitry ,  Marco Tagliasacchi

IPC: G10L19/035 ,  G06N20/00 ,  G10L19/038 ,  G10L25/18

Abstract: Systems and methods for training a machine-learned model are provided. A method can include can include obtaining an unlabeled audio signal, sampling the unlabeled audio signal to select one or more sampled slices, inputting the one or more sampled slices into a machine-learned model, receiving, as an output of the machine-learned model, one or more determined characteristics associated with the audio signal, determining a loss function for the machine-learned model based at least in part on a difference between the one or more determined characteristics and one or more corresponding ground truth characteristics of the audio signal, and training the machine-learned model from end to end based at least in part on the loss function. The one or more determined characteristics can include one or more reconstructed portions of the audio signal temporally adjacent to the one or more sampled slices or an estimated distance between two sampled slices.