公开(公告)号：US11676008B2

公开(公告)日：2023-06-13

申请号：US16577698

申请日：2019-09-20

Applicant: Google LLC

Inventor： Mark Sandler ,  Andrey Zhmoginov ,  Andrew Gerald Howard ,  Pramod Kaushik Mudrakarta

IPC: G10L15/16 ,  G06N3/08 ,  G06N3/045

CPC classification number: G06N3/08 ,  G06N3/045 ,  G10L15/16

Abstract: The present disclosure provides systems and methods that enable parameter-efficient transfer learning, multi-task learning, and/or other forms of model re-purposing such as model personalization or domain adaptation. In particular, as one example, a computing system can obtain a machine-learned model that has been previously trained on a first training dataset to perform a first task. The machine-learned model can include a first set of learnable parameters. The computing system can modify the machine-learned model to include a model patch, where the model patch includes a second set of learnable parameters. The computing system can train the machine-learned model on a second training dataset to perform a second task that is different from the first task, which may include learning new values for the second set of learnable parameters included in the model patch while keeping at least some (e.g., all) of the first set of parameters fixed.

公开(公告)号：US12169779B2

公开(公告)日：2024-12-17

申请号：US18310638

申请日：2023-05-02

Applicant: Google LLC

Inventor： Mark Sandler ,  Andrew Gerald Howard ,  Andrey Zhmoginov ,  Pramod Kaushik Mudrakarta

IPC: G10L15/16 ,  G06N3/045 ,  G06N3/08

Abstract: The present disclosure provides systems and methods that enable parameter-efficient transfer learning, multi-task learning, and/or other forms of model re-purposing such as model personalization or domain adaptation. In particular, as one example, a computing system can obtain a machine-learned model that has been previously trained on a first training dataset to perform a first task. The machine-learned model can include a first set of learnable parameters. The computing system can modify the machine-learned model to include a model patch, where the model patch includes a second set of learnable parameters. The computing system can train the machine-learned model on a second training dataset to perform a second task that is different from the first task, which may include learning new values for the second set of learnable parameters included in the model patch while keeping at least some (e.g., all) of the first set of parameters fixed.

公开(公告)号：US11734545B2

公开(公告)日：2023-08-22

申请号：US15898566

申请日：2018-02-17

Applicant: Google LLC

Inventor： Andrew Gerald Howard ,  Mark Sandler ,  Liang-Chieh Chen ,  Andrey Zhmoginov ,  Menglong Zhu

IPC: G06N3/04 ,  G06N3/08 ,  G06N3/045 ,  G06N3/048

CPC classification number: G06N3/04 ,  G06N3/045 ,  G06N3/08 ,  G06N3/048

Abstract: The present disclosure provides directed to new, more efficient neural network architectures. As one example, in some implementations, the neural network architectures of the present disclosure can include a linear bottleneck layer positioned structurally prior to and/or after one or more convolutional layers, such as, for example, one or more depthwise separable convolutional layers. As another example, in some implementations, the neural network architectures of the present disclosure can include one or more inverted residual blocks where the input and output of the inverted residual block are thin bottleneck layers, while an intermediate layer is an expanded representation. For example, the expanded representation can include one or more convolutional layers, such as, for example, one or more depthwise separable convolutional layers. A residual shortcut connection can exist between the thin bottleneck layers that play a role of an input and output of the inverted residual block.

公开(公告)号：US20240153116A1

公开(公告)日：2024-05-09

申请号：US17981891

申请日：2022-11-07

Applicant: Google LLC

Inventor： Kuntal Sengupta ,  Adarsh Kowdle ,  Andrey Zhmoginov ,  Hart Levy

IPC: G06T7/521 ,  G06F21/32 ,  G06V40/16 ,  G06V40/40

CPC classification number: G06T7/521 ,  G06F21/32 ,  G06V40/172 ,  G06V40/40 ,  G06T2207/10048 ,  G06T2207/20081 ,  G06T2207/30201

Abstract: Provided are computing systems, methods, and platforms for using machine-learned models to generate a depth map. The operations can include projecting, using a dot illuminator, near-infrared (NIR) dots on the scene. The NIR dots can have a uniform pattern. Additionally, the operations can include capturing, using a single NIR camera, the projected NIR dots on the scene. Moreover, the operations can include generating a dot image based on the captured NIR dots on the scene. Furthermore, the operations can include processing the dot image with a machine-learned model to generate a depth map of the scene. Subsequently, the operations can further include evaluating the generated depth map of the scene and a ground truth depth map, and performing an action based on the evaluation.

公开(公告)号：US11823024B2

公开(公告)日：2023-11-21

申请号：US17382503

申请日：2021-07-22

Applicant: Google LLC

Inventor： Andrew Gerald Howard ,  Mark Sandler ,  Liang-Chieh Chen ,  Andrey Zhmoginov ,  Menglong Zhu

IPC: G06N3/04 ,  G06N3/08 ,  G06N3/045 ,  G06N3/048

CPC classification number: G06N3/04 ,  G06N3/045 ,  G06N3/08 ,  G06N3/048

Abstract: The present disclosure provides directed to new, more efficient neural network architectures. As one example, in some implementations, the neural network architectures of the present disclosure can include a linear bottleneck layer positioned structurally prior to and/or after one or more convolutional layers, such as, for example, one or more depthwise separable convolutional layers. As another example, in some implementations, the neural network architectures of the present disclosure can include one or more inverted residual blocks where the input and output of the inverted residual block are thin bottleneck layers, while an intermediate layer is an expanded representation. For example, the expanded representation can include one or more convolutional layers, such as, for example, one or more depthwise separable convolutional layers. A residual shortcut connection can exist between the thin bottleneck layers that play a role of an input and output of the inverted residual block.

公开(公告)号：US20230267330A1

公开(公告)日：2023-08-24

申请号：US18310638

申请日：2023-05-02

Applicant: Google LLC

Inventor： Mark Sandler ,  Andrew Gerald Howard ,  Andrey Zhmoginov ,  Pramod Kaushik Mudrakarta

IPC: G06N3/08 ,  G06N3/045

CPC classification number: G06N3/08 ,  G06N3/045 ,  G10L15/16

Abstract: The present disclosure provides systems and methods that enable parameter-efficient transfer learning, multi-task learning, and/or other forms of model re-purposing such as model personalization or domain adaptation. In particular, as one example, a computing system can obtain a machine-learned model that has been previously trained on a first training dataset to perform a first task. The machine-learned model can include a first set of learnable parameters. The computing system can modify the machine-learned model to include a model patch, where the model patch includes a second set of learnable parameters. The computing system can train the machine-learned model on a second training dataset to perform a second task that is different from the first task, which may include learning new values for the second set of learnable parameters included in the model patch while keeping at least some (e.g., all) of the first set of parameters fixed.

公开(公告)号：US20230316081A1

公开(公告)日：2023-10-05

申请号：US18011873

申请日：2022-05-06

Applicant: Google LLC

Inventor： Mark Sandler ,  Andrey Zhmoginov ,  Thomas Edward Madams ,  Maksym Vladymyrov ,  Nolan Andrew Miller ,  Blaise Aguera-Arcas ,  Andrew Michael Jackson

IPC: G06N3/084 ,  G06N3/04

CPC classification number: G06N3/084 ,  G06N3/04

Abstract: The present disclosure provides a new type of generalized artificial neural network where neurons and synapses maintain multiple states. While classical gradient-based backpropagation in artificial neural networks can be seen as a special case of a two-state network where one state is used for activations and another for gradients with update rules derived from the chain rule, example implementations of the generalized framework proposed herein may additionally: have neither explicit notion of nor ever receive gradients; contain more than two states; and/or implement or apply learned (e.g., meta-learned) update rules that control updates to the state(s) of the neuron during forward and/or backward propagation of information.

公开(公告)号：US20190279092A1

公开(公告)日：2019-09-12

申请号：US16346313

申请日：2017-09-29

Applicant: Google LLC

Inventor： Mark Sandler ,  Andrey Zhmoginov ,  Soravit Changpinyo

IPC: G06N3/08 ,  G06N3/04

Abstract: Systems and methods of convolutional neural network compression are provided. For instance, a convolutional neural network can include an input convolutional layer having a plurality of associated input filters and an output convolutional layer having a plurality of associated output filters. The convolutional neural network implements a connection pattern defining connections between the plurality of input filters and the plurality of output filers. The connection pattern specifies that at least one output filter of the plurality of output filters is connected to only a subset of the plurality of input filters.

公开(公告)号：US20190147318A1

公开(公告)日：2019-05-16

申请号：US15898566

申请日：2018-02-17

Applicant: Google LLC

Inventor： Andrew Gerald Howard ,  Mark Sandler ,  Liang-Chieh Chen ,  Andrey Zhmoginov ,  Menglong Zhu

IPC: G06N3/04 ,  G06N3/08

Abstract: The present disclosure provides directed to new, more efficient neural network architectures. As one example, in some implementations, the neural network architectures of the present disclosure can include a linear bottleneck layer positioned structurally prior to and/or after one or more convolutional layers, such as, for example, one or more depthwise separable convolutional layers. As another example, in some implementations, the neural network architectures of the present disclosure can include one or more inverted residual blocks where the input and output of the inverted residual block are thin bottleneck layers, while an intermediate layer is an expanded representation. For example, the expanded representation can include one or more convolutional layers, such as, for example, one or more depthwise separable convolutional layers. A residual shortcut connection can exist between the thin bottleneck layers that play a role of an input and output of the inverted residual block.

公开(公告)号：US20250061333A1

公开(公告)日：2025-02-20

申请号：US18939318

申请日：2024-11-06

Applicant: Google LLC

Inventor： Mark Sandler ,  Andrew Gerald Howard ,  Andrey Zhmoginov ,  Pramod Kaushik Mudrakarta

IPC: G06N3/08 ,  G06N3/045 ,  G10L15/16

Abstract: The present disclosure provides systems and methods that enable parameter-efficient transfer learning, multi-task learning, and/or other forms of model re-purposing such as model personalization or domain adaptation. In particular, as one example, a computing system can obtain a machine-learned model that has been previously trained on a first training dataset to perform a first task. The machine-learned model can include a first set of learnable parameters. The computing system can modify the machine-learned model to include a model patch, where the model patch includes a second set of learnable parameters. The computing system can train the machine-learned model on a second training dataset to perform a second task that is different from the first task, which may include learning new values for the second set of learnable parameters included in the model patch while keeping at least some (e.g., all) of the first set of parameters fixed.