Generating parameter values for recurrent neural networks

    公开(公告)号:US11900235B1

    公开(公告)日:2024-02-13

    申请号:US17470716

    申请日:2021-09-09

    Applicant: Google LLC

    CPC classification number: G06N3/04 G06N3/08

    Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for processing inputs using recurrent neural networks. One of the systems includes a main recurrent neural network comprising one or more recurrent neural network layers and a respective hyper recurrent neural network corresponding to each of the one or more recurrent neural network layers, wherein each hyper recurrent neural network is configured to, at each of a plurality of time steps: process the layer input at the time step to the corresponding recurrent neural network layer, the current layer hidden state of the corresponding recurrent neural network layer, and a current hypernetwork hidden state of the hyper recurrent neural network to generate an updated hypernetwork hidden state.

    Learning longer-term dependencies in neural network using auxiliary losses

    公开(公告)号:US11501168B2

    公开(公告)日:2022-11-15

    申请号:US16273041

    申请日:2019-02-11

    Applicant: Google LLC

    Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for structuring and training a recurrent neural network. This describes a technique that improves the ability to capture long term dependencies in recurrent neural networks by adding an unsupervised auxiliary loss at one or more anchor points to the original objective. This auxiliary loss forces the network to either reconstruct previous events or predict next events in a sequence, making truncated backpropagation feasible for long sequences and also improving full backpropagation through time.

    TRAINING NEURAL NETWORKS USING LEARNED ADAPTIVE LEARNING RATES

    公开(公告)号:US20210034973A1

    公开(公告)日:2021-02-04

    申请号:US16943957

    申请日:2020-07-30

    Applicant: Google LLC

    Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for training a neural network. One of the methods includes training the neural network for one or more training steps in accordance with a current learning rate; generating a training dynamics observation characterizing the training of the trainee neural network on the one or more training steps; providing the training dynamics observation as input to a controller neural network that is configured to process the training dynamics observation to generate a controller output that defines an updated learning rate; obtaining as output from the controller neural network the controller output that defines the updated learning rate; and setting the learning rate to the updated learning rate.

    Processing clinical notes using recurrent neural networks

    公开(公告)号:US10770180B1

    公开(公告)日:2020-09-08

    申请号:US16712947

    申请日:2019-12-12

    Applicant: Google LLC

    Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for predicting future patient health using neural networks. One of the methods includes receiving electronic health record data for a patient; generating a respective observation embedding for each of the observations, comprising, for each clinical note: processing the sequence of tokens in the clinical note using a clinical note embedding LSTM to generate a respective token embedding for each of the tokens; and generating the observation embedding for the clinical note from the token embeddings; generating an embedded representation, comprising, for each time window: combining the observation embeddings of observations occurring during the time window to generate a patient record embedding; and processing the embedded representation of the electronic health record data using a prediction recurrent neural network to generate a neural network output that characterizes a future health status of the patient.

    Co-Training of Action Recognition Machine Learning Models

    公开(公告)号:US20250037426A1

    公开(公告)日:2025-01-30

    申请号:US18716912

    申请日:2022-12-09

    Applicant: Google LLC

    Abstract: A method includes obtaining video datasets each including pairs of a training video and a ground-truth action classification of the training video. The method also includes generating an action recognition model that includes a shared encoder model and action classification heads. A number of the action classifications heads may be equal to a number of the video datasets, and each action classification head may be configured to, based on an output of the shared encoder model, classify training videos sampled from a corresponding video dataset. The method also includes determining, by the action recognition model and for each training video sampled from the video datasets, an inferred action classification. The method further includes determining a loss value based on the inferred action classifications and the ground-truth action classifications, and adjusting parameters of the action recognition model based on the loss value.

    Training a document classification neural network

    公开(公告)号:US11200492B1

    公开(公告)日:2021-12-14

    申请号:US16735453

    申请日:2020-01-06

    Applicant: Google LLC

    Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for training a document classification neural network. One of the methods includes training an autoencoder neural network to autoencode input documents, wherein the autoencoder neural network comprises the one or more LSTM neural network layers and an autoencoder output layer, and wherein training the autoencoder neural network comprises determining pre-trained values of the parameters of the one or more LSTM neural network layers from initial values of the parameters of the one or more LSTM neural network layers; and training the document classification neural network on a plurality of training documents to determine trained values of the parameters of the one or more LSTM neural network layers from the pre-trained values of the parameters of the one or more LSTM neural network layers.

Patent Agency Ranking