中科微点-全球专利检索

  1. Login
  2. Sign Up

Search Results

16 records (took 0.014 seconds)

    Generating trained neural networks with increased robustness against adversarial attacks
    3.
    发明授权
    Generating trained neural networks with increased robustness against adversarial attacks 有权

    公开(公告)号：US11481617B2

    公开(公告)日：2022-10-25

    申请号：US16253561

    申请日：2019-01-22

    Applicant: Adobe Inc.

    Inventor： Mayank Singh ,  Nupur Kumari ,  Dhruv Khattar ,  Balaji Krishnamurthy ,  Abhishek Sinha

    IPC: G06N3/08 ,  G06N20/00 ,  H04L9/40

    Abstract: The present disclosure relates to systems, methods, and non-transitory computer readable media for generating trained neural network with increased robustness against adversarial attacks by utilizing a dynamic dropout routine and/or a cyclic learning rate routine. For example, the disclosed systems can determine a dynamic dropout probability distribution associated with neurons of a neural network. The disclosed systems can further drop neurons from a neural network based on the dynamic dropout probability distribution to help neurons learn distinguishable features. In addition, the disclosed systems can utilize a cyclic learning rate routine to force copy weights of a copy neural network away from weights of an original neural network without decreasing prediction accuracy to ensure that the decision boundaries learned are different.

    GENERATING TRAINED NEURAL NETWORKS WITH INCREASED ROBUSTNESS AGAINST ADVERSARIAL ATTACKS
    4.
    发明申请
    GENERATING TRAINED NEURAL NETWORKS WITH INCREASED ROBUSTNESS AGAINST ADVERSARIAL ATTACKS 审中-公开

    公开(公告)号：US20200234110A1

    公开(公告)日：2020-07-23

    申请号：US16253561

    申请日：2019-01-22

    Applicant: Adobe Inc.

    Inventor： Mayank Singh ,  Nupur Kumari ,  Dhruv Khattar ,  Balaji Krishnamurthy ,  Abhishek Sinha

    IPC: G06N3/08 ,  H04L29/06 ,  G06N20/00

    Abstract: The present disclosure relates to systems, methods, and non-transitory computer readable media for generating trained neural network with increased robustness against adversarial attacks by utilizing a dynamic dropout routine and/or a cyclic learning rate routine. For example, the disclosed systems can determine a dynamic dropout probability distribution associated with neurons of a neural network. The disclosed systems can further drop neurons from a neural network based on the dynamic dropout probability distribution to help neurons learn distinguishable features. In addition, the disclosed systems can utilize a cyclic learning rate routine to force copy weights of a copy neural network away from weights of an original neural network without decreasing prediction accuracy to ensure that the decision boundaries learned are different.

    Digital Image Search Training using Aggregated Digital Images
    5.
    发明申请
    Digital Image Search Training using Aggregated Digital Images 审中-公开

    公开(公告)号：US20200134056A1

    公开(公告)日：2020-04-30

    申请号：US16177243

    申请日：2018-10-31

    Applicant: Adobe Inc.

    Inventor： Mausoom Sarkar ,  Hiresh Gupta ,  Abhishek Sinha

    IPC: G06F17/30 ,  G06N3/08

    Abstract: Digital image search training techniques and machine-learning architectures are described. In one example, a query digital image is received by service provider system, which is then used to select at least one positive sample digital image, e.g., having a same product ID. A plurality of negative sample digital images is also selected by the service provider system based on the query digital image, e.g., having different product IDs. The at least one positive sample digital image and the plurality of negative samples are then aggregated by the service provider system into a single aggregated digital image. At least one neural network is then trained by the service provider system using a loss function based on a feature comparison between the query digital image and samples from the aggregated digital image in a single pass.

    Generating trained neural networks with increased robustness against adversarial attacks
    6.
    发明授权
    Generating trained neural networks with increased robustness against adversarial attacks 有权

    公开(公告)号：US11829880B2

    公开(公告)日：2023-11-28

    申请号：US18049209

    申请日：2022-10-24

    Applicant: Adobe Inc.

    Inventor： Mayank Singh ,  Nupur Kumari ,  Dhruv Khattar ,  Balaji Krishnamurthy ,  Abhishek Sinha

    IPC: G06N3/08 ,  G06N20/00 ,  H04L9/40

    CPC classification number: G06N3/08 ,  G06N20/00 ,  H04L63/1441

    Abstract: The present disclosure relates to systems, methods, and non-transitory computer readable media for generating trained neural network with increased robustness against adversarial attacks by utilizing a dynamic dropout routine and/or a cyclic learning rate routine. For example, the disclosed systems can determine a dynamic dropout probability distribution associated with neurons of a neural network. The disclosed systems can further drop neurons from a neural network based on the dynamic dropout probability distribution to help neurons learn distinguishable features. In addition, the disclosed systems can utilize a cyclic learning rate routine to force copy weights of a copy neural network away from weights of an original neural network without decreasing prediction accuracy to ensure that the decision boundaries learned are different.

    IDENTIFYING DIGITAL ATTRIBUTES FROM MULTIPLE ATTRIBUTE GROUPS WITHIN TARGET DIGITAL IMAGES UTILIZING A DEEP COGNITIVE ATTRIBUTION NEURAL NETWORK
    7.
    发明申请
    IDENTIFYING DIGITAL ATTRIBUTES FROM MULTIPLE ATTRIBUTE GROUPS WITHIN TARGET DIGITAL IMAGES UTILIZING A DEEP COGNITIVE ATTRIBUTION NEURAL NETWORK 有权

    公开(公告)号：US20210073267A1

    公开(公告)日：2021-03-11

    申请号：US16564831

    申请日：2019-09-09

    Applicant: Adobe, Inc.

    Inventor： Ayush Chopra ,  Mausoom Sarkar ,  Jonas Dahl ,  Hiresh Gupta ,  Balaji Krishnamurthy ,  Abhishek Sinha

    IPC: G06F16/535 ,  G06K9/62 ,  G06F16/55 ,  G06N3/04 ,  G06F17/15

    Abstract: The present disclosure relates to systems, methods, and non-transitory computer-readable media for generating tags for an object portrayed in a digital image based on predicted attributes of the object. For example, the disclosed systems can utilize interleaved neural network layers of alternating inception layers and dilated convolution layers to generate a localization feature vector. Based on the localization feature vector, the disclosed systems can generate attribute localization feature embeddings, for example, using some pooling layer such as a global average pooling layer. The disclosed systems can then apply the attribute localization feature embeddings to corresponding attribute group classifiers to generate tags based on predicted attributes. In particular, attribute group classifiers can predict attributes as associated with a query image (e.g., based on a scoring comparison with other potential attributes of an attribute group). Based on the generated tags, the disclosed systems can respond to tag queries and search queries.

    Systems and methods of training neural networks against adversarial attacks
    8.
    发明授权
    Systems and methods of training neural networks against adversarial attacks 有权

    公开(公告)号：US11734565B2

    公开(公告)日：2023-08-22

    申请号：US17805405

    申请日：2022-06-03

    Applicant: ADOBE INC.

    Inventor： Mayank Singh ,  Abhishek Sinha ,  Balaji Krishnamurthy

    IPC: G06N3/08 ,  G06F21/57 ,  G06N5/046 ,  G06F18/24 ,  G06F18/214 ,  G06N3/048

    CPC classification number: G06N3/08 ,  G06F18/214 ,  G06F18/24 ,  G06F21/577 ,  G06N3/048 ,  G06N5/046

    Abstract: Embodiments disclosed herein describe systems, methods, and products that generate trained neural networks that are robust against adversarial attacks. During a training phase, an illustrative computer may iteratively optimize a loss function that may include a penalty for ill-conditioned weight matrices in addition to a penalty for classification errors. Therefore, after the training phase, the trained neural network may include one or more well-conditioned weight matrices. The one or more well-conditioned weight matrices may minimize the effect of perturbations within an adversarial input thereby increasing the accuracy of classification of the adversarial input. By contrast, conventional training approaches may merely reduce the classification errors using backpropagation, and, as a result, any perturbation in an input is prone to generate a large effect on the output.

Patent Agency Ranking