中科微点-全球专利检索

  1. Login
  2. Sign Up

Search Results

812 records (took 0.037 seconds)

    MODEL-AWARE METHOD AND SYSTEM FOR TRAINING AND/OR FINE-TUNING A MACHINE LEARNING MODEL
    121.
    发明公开
    MODEL-AWARE METHOD AND SYSTEM FOR TRAINING AND/OR FINE-TUNING A MACHINE LEARNING MODEL 审中-公开

    公开(公告)号：US20230289616A1

    公开(公告)日：2023-09-14

    申请号：US18319636

    申请日：2023-05-18

    Applicant: Lemon Inc.

    Inventor： Tongping Liu ,  Wei Xu ,  Jianjun Chen

    IPC: G06N3/098

    CPC classification number: G06N3/098

    Abstract: System and method of training a machine learning model on a plurality of devices in parallel are provided. The method includes performing a model profiling execution before a model normal execution, allocating tensors of the model into a plurality of chunks based on profiling results from the model profiling execution, and performing the model normal execution on the plurality of devices in parallel to train or fine-tune the model.

    CREATING EFFECT ASSETS WHILE AVOIDING SIZE INFLATION
    122.
    发明公开
    CREATING EFFECT ASSETS WHILE AVOIDING SIZE INFLATION 审中-公开

    公开(公告)号：US20230281162A1

    公开(公告)日：2023-09-07

    申请号：US17687461

    申请日：2022-03-04

    Applicant: Lemon Inc.

    Inventor： Zeyong Cai ,  Nite Luo

    IPC: G06F16/13 ,  G06F16/16 ,  G06F16/955 ,  G06F9/445

    CPC classification number: G06F16/13 ,  G06F9/445 ,  G06F16/168 ,  G06F16/9566

    Abstract: The present disclosure describes techniques for effect asset creation. At one file of defining at least one new type of asset may be created based on an existing type of asset. Properties of the at least one new type of asset may be configured. The properties comprise an identifier of the at least one new type of asset and information indicative of the existing type of asset. The at least one new type of asset may be implemented with scripts. Implementing the at least one new type of asset with scripts comprises fetching a native object corresponding to the existing type of asset. The script-based at least one new type of asset enables to create new effect assets while avoiding an inflation of a package size of an effect creation tool.

    BOX DETECTION FOR OBJECT ATTACHMENT
    124.
    发明公开
    BOX DETECTION FOR OBJECT ATTACHMENT 审中-公开

    公开(公告)号：US20230252752A1

    公开(公告)日：2023-08-10

    申请号：US17666045

    申请日：2022-02-07

    Applicant: Lemon Inc.

    Inventor： Shuo Cheng ,  Peng Wang

    IPC: G06V10/25 ,  G06T7/50 ,  G06T7/70 ,  G06N3/04

    CPC classification number: G06V10/25 ,  G06N3/0454 ,  G06T7/50 ,  G06T7/70 ,  G06T2207/20076

    Abstract: The present disclosure describes techniques for determining a bounding box. An image may be received. An X-frame, a Y-frame, and a normal frame may be estimated based on the image using a first neural network. At least one planar region may be detected from the image using a second neural network. A vanishing point detection may be performed on each of the at least one planar region. Output of the first neural network may be fused with results of the vanishing point detection. A depth value of each pixel in at least one plane corresponding to the at least one planar region may be determined based at least in part on a result of the fusing. A location of a bounding box may be determined based at least in part on the depth value of each pixel in the at least one plane.

    FLEXIBLE LOAD BALANCING ON MULTIPATH NETWORKS
    125.
    发明公开
    FLEXIBLE LOAD BALANCING ON MULTIPATH NETWORKS 审中-公开

    公开(公告)号：US20230246966A1

    公开(公告)日：2023-08-03

    申请号：US18295783

    申请日：2023-04-04

    Applicant: Lemon Inc.

    Inventor： Yan SUN ,  Hang ZHU ,  Lishan LI ,  Jianxi YE ,  Huan SUN

    IPC: H04L47/125 ,  H04L45/24

    CPC classification number: H04L47/125 ,  H04L45/24

    Abstract: A computer system for flexible load balancing on a multipath network includes a processor that implements a multipath transport protocol as a transport layer of a network stack, a load balancer that distributes network traffic across a plurality of paths, and a congestion controller in communication with the load balancer. The congestion controller determines parameters for a message based on information received from the load balancer. A scheduler included in the load balancer selects a load balancing algorithm from a plurality of load balancing algorithms based on the parameters of the message received from the congestion controller and, based on the selected load balancing algorithm, determines a timing and a path for the message to be sent to the transport layer.

    IMPLEMENTATION OF INSTANT CORRUPTION DETECTION AND RECOVERY
    128.
    发明公开
    IMPLEMENTATION OF INSTANT CORRUPTION DETECTION AND RECOVERY 审中-公开

    公开(公告)号：US20230237046A1

    公开(公告)日：2023-07-27

    申请号：US18130670

    申请日：2023-04-04

    Applicant: Lemon Inc. ,  Beijing Youzhuju Network Technology Co., Ltd.

    Inventor： Viacheslav Dubeyko ,  Jian Wang

    IPC: G06F16/23 ,  G06F16/2455

    CPC classification number: G06F16/2365 ,  G06F16/24568

    Abstract: The present disclosure describes techniques for implementing instant corruption detection and recovery. A plurality of streams may be created in a storage device. Each of the plurality of streams may contain a sequence of metadata nodes of a same type. Each of the plurality of streams may maintain an initial state, a sequence of delta modifications to the initial state, and an actual state for each of the sequence of metadata nodes. A checking and recovery function associated with a particular stream among the plurality of streams may be determined. The checking and recovery function may comprise checking logic configured to detect corruptions by checking modification operations associated with metadata nodes in the particular stream. The checking and recovery function may further comprise recovery logic configured to perform recoveries from the corruptions. The checking and recovery function associated with the particular stream may be implemented in the storage device.

    EMBEDDING OPTIMIZATION FOR A MACHINE LEARNING MODEL
    129.
    发明公开
    EMBEDDING OPTIMIZATION FOR A MACHINE LEARNING MODEL 审中-公开

    公开(公告)号：US20230229736A1

    公开(公告)日：2023-07-20

    申请号：US17579566

    申请日：2022-01-19

    Applicant: Lemon Inc.

    Inventor： Xia XIAO ,  Ming CHEN ,  Youlong CHENG

    IPC: G06K9/62 ,  G06F17/16 ,  G06N20/00

    CPC classification number: G06K9/6257 ,  G06F17/16 ,  G06N20/00

    Abstract: Embodiments of the present disclosure relate to feature selection via an ensemble of gating layers. According to embodiments of the present disclosure, a set of model parameter values for a machine learning model and a set of embedding vectors are determined for an input field of the machine learning model. The machine learning model is constructed to map an input sample in the input field to an embedding vector in the embedding vectors and process the embedding vector with the model parameter values to generate a model output. The machine learning model is trained by updating the model parameter values and the embedding vectors according to at least a first training objective function, the first training objective function being based on an orthogonality metric between embedding vectors in the embedding vectors and based on a difference between the model output and a ground-truth model output.

    Accessing user accounts and data from any computing device
    130.
    发明授权
    Accessing user accounts and data from any computing device 有权

    公开(公告)号：US11693986B1

    公开(公告)日：2023-07-04

    申请号：US17877662

    申请日：2022-07-29

    Applicant: Lemon Inc. ,  Beijing Youzhuju Network Technology Co., Ltd.

    Inventor： Viacheslav Dubeyko ,  Jian Wang

    IPC: G06F21/62 ,  H04L9/40 ,  G06F21/60 ,  H04L29/06

    CPC classification number: G06F21/6218 ,  G06F21/604 ,  H04L63/0838 ,  H04L63/0861 ,  H04L63/105

    Abstract: The present disclosure describes techniques for accessing user accounts and data from any computing device. It may be determined whether an account of a user exists in a cloud service in response to receiving information associated with the user from any computing device. Data associated with the account may be stored by the cloud service. There may be a plurality of types of data associated with a plurality of security levels. The plurality of security levels may correspond to different security requirements. The data associated with the account may belong to at least one of the plurality of types of data. An instance of the account may be deployed to the computing device in response to determining that the account exists in the cloud service. The instance of the account may enable the user to access services via the computing device.

Patent Agency Ranking