Training a character through interactions

    公开(公告)号:US12056581B1

    公开(公告)日:2024-08-06

    申请号:US16786097

    申请日:2020-02-10

    Applicant: Apple Inc.

    CPC classification number: G06N20/00 G06N3/006 G06N5/022 G06N5/04 G06N7/01

    Abstract: Various implementations disclosed herein include devices, systems, and methods for training of an action determining component of a computer character. In some implementations, actions are taken by the character in a 3D environment according to an action determining component of the character, where the character is rewarded or penalized for interactions associated with an object/concept in the 3D environment according to an assigned object/concept reward or penalty. In some implementations, the reward or the penalty assigned to the object/concept is modified, and the character is then rewarded or penalized for interactions associated with the object/concept according to the modified reward or the modified penalty. The action determining component of the character is trained using a reinforcement learning technique that accounts for rewards or penalties obtained by virtual character for interactions associated with the object/concept.

    Procedural generation of computer objects

    公开(公告)号:US11182980B1

    公开(公告)日:2021-11-23

    申请号:US17169988

    申请日:2021-02-08

    Applicant: Apple Inc.

    Abstract: Various implementations disclosed herein include devices, systems, and methods for identifying virtual object kits that may be used to generate a target virtual object. In some implementations, a method includes obtaining a plurality of virtual object kits that generate variations of respective virtual object types. Each virtual object kit is associated with a characteristic parameter. For each virtual object kit, a subset of possible virtual object variations of the virtual object type is generated by assigning corresponding values to the characteristic parameter based on a distribution criterion. A subset of the plurality of virtual object kits is identified based on the target virtual object being within a similarity threshold of one or more virtual object variations generated using the subset of the plurality of virtual object kits in response to a request to assemble a target virtual object.

Patent Agency Ranking