Neural network optimization
    92.
    发明授权

    公开(公告)号:US11853873B2

    公开(公告)日:2023-12-26

    申请号:US16152348

    申请日:2018-10-04

    CPC classification number: G06N3/08 G06N3/04

    Abstract: A method of reducing kernel computations; the method comprising ordering a plurality of kernel channels. A first of the ordered kernel channels is then convolved with input data to produce a convolution output, and it is determined whether to convolve one or more subsequent kernel channels of the ordered kernel channels. Determining whether to convolve subsequent kernel channels comprises considering a potential contribution of at least one of the one or more subsequent kernel channels in combination with the convolution output.

    Method and apparatus for data processing operation

    公开(公告)号:US11809836B2

    公开(公告)日:2023-11-07

    申请号:US17005046

    申请日:2020-08-27

    Applicant: Arm Limited

    CPC classification number: G06F7/5443 G06F16/258 G06F18/213 G06N3/04 H03M7/24

    Abstract: A system includes a fixed-point accumulator for storing numbers in an anchored fixed-point number format, a data interface arranged to receive a plurality of weight values and a plurality of data values represented in a floating-point number format, and logic circuitry. The logic circuitry is configured to: determine an anchor value indicative of a value of a lowest significant bit of the anchored fixed-point number format; convert at least a portion of the plurality of data values to the anchored fixed-point number format; perform MAC operations between the converted at least portion and respective weight values, using fixed-point arithmetic, to generate an accumulation value in the anchored fixed-point number format; and determine an output element of a later of a neural network in dependence on the accumulation value.

    DATA PROCESSING SYSTEMS
    94.
    发明公开

    公开(公告)号:US20230316694A1

    公开(公告)日:2023-10-05

    申请号:US17713119

    申请日:2022-04-04

    Applicant: Arm Limited

    CPC classification number: G06V10/25 G06N3/0454 G06V10/82 G06V10/87

    Abstract: When processing regions of interest in frames in a data processing system that can execute a plurality of neural networks that are each configured to more optimally process a region of interest of a respective size, a region of interest is first identified within a frame, the size of the region of interest is determined, and one of the plurality of available neural networks is selected to use to process the region of interest based on the determined size. The region of interest is scaled to produce a scaled version of the region of interest, where the scaling is determined based on the selected neural network. The scaled version of the region of interest is then processed using the selected neural network.

    Augmented reality system
    95.
    发明授权

    公开(公告)号:US11670157B2

    公开(公告)日:2023-06-06

    申请号:US17339710

    申请日:2021-06-04

    Applicant: Arm Limited

    Abstract: An AR system includes a user interface, one or more sensors arranged to generate sensor data representing part of an environment in which a user of the AR system is located, and a memory. The memory is arranged to store object association data associating the user with one or more objects in the environment, and object location data indicating a respective location of each of the one or more objects. The AR system is arranged to determine a position of the user; determine an updated location of one of the one or more objects in dependence on the generated sensor data and the determined position of the user, update the stored object location data to indicate the determined updated location of said one of the one or more objects, and output information depending on the updated location of said one of the one or more objects via the user interface.

    Apparatus and Method of Focusing Light

    公开(公告)号:US20230115678A1

    公开(公告)日:2023-04-13

    申请号:US17484490

    申请日:2021-09-24

    Applicant: Arm Limited

    Abstract: There is provided a display apparatus to focus light for a user. The apparatus comprises a tuneable lens having controllable optical properties, an eye-tracker device to determine a position at which the user is looking, and circuitry to control the optical properties of the tuneable lens to bring an object at the depth of the position into focus for the user. A method of focusing light is also provided. The method comprises determining a position at which the user is looking and controlling optical properties of a tuneable lens to bring an object at the depth of the position into focus for the user.

    Human-machine interface system
    100.
    发明授权

    公开(公告)号:US11513597B2

    公开(公告)日:2022-11-29

    申请号:US17211107

    申请日:2021-03-24

    Applicant: Arm Limited

    Abstract: A human-machine interface system comprises a sensor configured to generate data associated with a human movement, such as measured electrical signals or data from an accelerometer. A measurement unit of the human-machine interface measures user movement over time to generate a sequence of measured user movement data. A processor processes the data associated with a human movement from the sensor using a trained neural network to determine one or more predicted user actions. A comparison unit compares the one or more predicted user actions with one or more user actions obtained from the sequence of measured user movement data. A control unit uses the predicted user actions to control a process in an information processing apparatus in dependence upon the comparison performed by the comparison unit.

Patent Agency Ranking