Using embedding functions with a deep network

    公开(公告)号:US11481631B1

    公开(公告)日:2022-10-25

    申请号:US16895855

    申请日:2020-06-08

    Applicant: Google LLC

    Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for using embedded function with a deep network. One of the methods includes receiving an input comprising a plurality of features, wherein each of the features is of a different feature type; processing each of the features using a respective embedding function to generate one or more numeric values, wherein each of the embedding functions operates independently of each other embedding function, and wherein each of the embedding functions is used for features of a respective feature type; processing the numeric values using a deep network to generate a first alternative representation of the input, wherein the deep network is a machine learning model composed of a plurality of levels of non-linear operations; and processing the first alternative representation of the input using a logistic regression classifier to predict a label for the input.

    Secure conversion tracking
    22.
    发明授权

    公开(公告)号:US10963891B2

    公开(公告)日:2021-03-30

    申请号:US15666378

    申请日:2017-08-01

    Applicant: Google LLC

    Abstract: In response to a conversion action associated with an online advertisement (“ad”), an advertiser associated with the ad generates a unique identifier (e.g., a pseudo random number), which is transmitted to a payment system through a user device. In some implementations, the unique identifier can be generated by a payment system in response to a request from an advertiser. The advertiser stores the unique identifier in a secure location that is accessible by the payment system. The payment system compares the received unique identifier with the stored unique identifier. If the received unique identifier matches the stored unique identifier, and if the received unique identifier was not previously generated by the advertiser, then the payment system deems the conversion action to be valid.

    CONVERSATIONAL TURN ANALYSIS NEURAL NETWORKS
    24.
    发明申请

    公开(公告)号:US20190294973A1

    公开(公告)日:2019-09-26

    申请号:US16363891

    申请日:2019-03-25

    Applicant: Google LLC

    Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for training conversational turn analysis neural networks. One of the methods includes obtaining unsupervised training data comprising a plurality of dialogue transcripts; training a turn prediction neural network to perform a turn prediction task on the unsupervised training data using unsupervised learning, wherein: the turn prediction neural network comprises (i) a turn encoder neural network and (ii) a turn decoder neural network; obtaining supervised training data; and training a supervised prediction neural network to perform a supervised prediction task on the supervised training data using supervised learning.

Patent Agency Ranking