知搜-全球专利检索

  1. 登录
  2. 注册

搜索结果

14 条记录 (耗时 0.017 秒)

    Music modeling
    2.
    发明授权
    Music modeling 有权

    公开(公告)号：US09792889B1

    公开(公告)日：2017-10-17

    申请号：US15343059

    申请日：2016-11-03

    申请人： INTERNATIONAL BUSINESS MACHINES CORPORATION

    发明人： Yachiko Obara ,  Shohei Ohsawa ,  Takayuki Osogami

    IPC分类号： G04B13/00 ,  G10H1/00

    CPC分类号： G10H1/0025 ,  G10H2210/061 ,  G10H2210/066 ,  G10H2210/145 ,  G10H2240/145 ,  G10H2250/005 ,  G10H2250/311

    摘要： A computer implemented method is provided for generating a prediction of a next musical note by a computer having at least a processor and a memory. A computer processor system is also provided for generating a prediction of a next musical note. The method includes storing sequential musical notes in the memory. The method further includes generating, by the processor, the prediction of the next musical note based upon a music model and the sequential musical notes stored in the memory. The method also includes updating, by the processor, the music model based upon the prediction of the next musical note and an actual one of the next musical note. The method additionally includes resetting, by the processor, the memory at fixed time intervals.

    CONTINUOUS CONTROL OF ATTENTION FOR A DEEP LEARNING NETWORK
    4.
    发明申请
    CONTINUOUS CONTROL OF ATTENTION FOR A DEEP LEARNING NETWORK 审中-公开

    公开(公告)号：US20200026247A1

    公开(公告)日：2020-01-23

    申请号：US16039934

    申请日：2018-07-19

    申请人： INTERNATIONAL BUSINESS MACHINES CORPORATION

    发明人： Shohei Ohsawa ,  Takayuki Osogami

    IPC分类号： G05B13/02 ,  G06N3/04 ,  G06N7/08

    摘要： A computer-implemented method for reducing computation cost associated with a machine learning task performed by a computer system by implementing continuous control of attention for a deep learning network includes initializing a control-value function, an observation-value function and a sequence of states associated with a current episode. If a current epoch associated with the current episode is odd, an observation-action is selected, the observation-action is executed to observe a partial image, and the observation-value function is updated based on the partial image and the control-value function. If the current epoch is even, a control-action is selected, the control-action is executed to obtain a reward corresponding to the control-action, and the control-value function is updated based on the reward and the observation-value function.

    Rating model generation
    6.
    发明授权
    Rating model generation 有权

    公开(公告)号：US11093846B2

    公开(公告)日：2021-08-17

    申请号：US15201068

    申请日：2016-07-01

    申请人： INTERNATIONAL BUSINESS MACHINES CORPORATION

    发明人： Yachiko Obara ,  Shohei Ohsawa ,  Takayuki Osogami

    IPC分类号： G06N7/00 ,  G06F17/16 ,  G06Q30/06 ,  G06N20/00 ,  G06F16/335

    摘要： Rating models may be generated by obtaining a plurality of consumption values, obtaining a plurality of rating values, training a model that estimates consumption values and rating values by utilizing a plurality of consumer attributes for each consumer, a plurality of item attributes for each item, and a plurality of weights for each attribute of each combination of a consumer and an item. Each estimated consumption value is a function of the plurality of weights for each attribute of each combination of each consumer and each item that corresponds with the estimated consumption value, and each estimated rating value is a function of the plurality of consumer attributes of a consumer that corresponds with the estimated rating value, the plurality of item attributes of an item that corresponds with the estimated rating value, and the plurality of weights that corresponds with the estimated rating value.

    Reducing computational costs of deep reinforcement learning by gated convolutional neural network
    8.
    发明授权
    Reducing computational costs of deep reinforcement learning by gated convolutional neural network 审中-公开

    公开(公告)号：US10671891B2

    公开(公告)日：2020-06-02

    申请号：US16039679

    申请日：2018-07-19

    申请人： INTERNATIONAL BUSINESS MACHINES CORPORATION

    发明人： Shohei Ohsawa ,  Takayuki Osogami

    IPC分类号： G06K9/00 ,  G06K9/62 ,  G06N3/08 ,  G06N3/04

    摘要： A method is provided for reducing a computational cost of deep reinforcement learning using an input image to provide a filtered output image composed of pixels. The method includes generating a moving gate in which the pixels of the filtered output image to be masked are assigned a first gate value and the pixels of the filtered output image to be passed through are assigned a second gate value. The method further includes applying the input image and the moving gate to a GCNN to provide the filtered output image such that only the pixels of the input image used to compute the pixels assigned the second gate value are processed by the GCNN while bypassing the pixels of the input image useable to compute the pixels assigned the first gate to reduce an overall processing time of the input image in order to provide the filtered output image.

    Continuous control of attention for a deep learning network
    9.
    发明授权
    Continuous control of attention for a deep learning network 有权

    公开(公告)号：US11188035B2

    公开(公告)日：2021-11-30

    申请号：US16039934

    申请日：2018-07-19

    申请人： INTERNATIONAL BUSINESS MACHINES CORPORATION

    发明人： Shohei Ohsawa ,  Takayuki Osogami

    IPC分类号： G06N3/04 ,  G05B13/02 ,  G06N7/08

    摘要： A computer-implemented method for reducing computation cost associated with a machine learning task performed by a computer system by implementing continuous control of attention for a deep learning network includes initializing a control-value function, an observation-value function and a sequence of states associated with a current episode. If a current epoch associated with the current episode is odd, an observation-action is selected, the observation-action is executed to observe a partial image, and the observation-value function is updated based on the partial image and the control-value function. If the current epoch is even, a control-action is selected, the control-action is executed to obtain a reward corresponding to the control-action, and the control-value function is updated based on the reward and the observation-value function.

    REDUCING COMPUTATIONAL COSTS OF DEEP REINFORCEMENT LEARNING BY GATED CONVOLUTIONAL NEURAL NETWORK
    10.
    发明申请
    REDUCING COMPUTATIONAL COSTS OF DEEP REINFORCEMENT LEARNING BY GATED CONVOLUTIONAL NEURAL NETWORK 审中-公开

    公开(公告)号：US20200026963A1

    公开(公告)日：2020-01-23

    申请号：US16039679

    申请日：2018-07-19

    申请人： INTERNATIONAL BUSINESS MACHINES CORPORATION

    发明人： Shohei Ohsawa ,  Takayuki Osogami

    IPC分类号： G06K9/62 ,  G06K9/00 ,  G06N3/04 ,  G06N3/08

    摘要： A method is provided for reducing a computational cost of deep reinforcement learning using an input image to provide a filtered output image composed of pixels. The method includes generating a moving gate in which the pixels of the filtered output image to be masked are assigned a first gate value and the pixels of the filtered output image to be passed through are assigned a second gate value. The method further includes applying the input image and the moving gate to a GCNN to provide the filtered output image such that only the pixels of the input image used to compute the pixels assigned the second gate value are processed by the GCNN while bypassing the pixels of the input image useable to compute the pixels assigned the first gate to reduce an overall processing time of the input image in order to provide the filtered output image.