公开(公告)号：US12052315B2

公开(公告)日：2024-07-30

申请号：US17129579

申请日：2020-12-21

Applicant: Apple Inc.

Inventor： Stephen Cosman ,  Kalu Onuka Kalu ,  Marcelo Lotif Araujo ,  Michael Chatzidakis ,  Thi Hai Van Do ,  Alexis Hugo Louis Durocher ,  Guillaume Tartavel ,  Sowmya Gopalan ,  Vignesh Jagadeesh ,  Abhishek Bhowmick ,  John Duchi ,  Julien Freudiger ,  Gaurav Kapoor ,  Ryan M. Rogers

IPC: H04L67/1097 ,  G06F16/2457 ,  G06F16/438 ,  G06F16/44 ,  G06F18/214 ,  G06F21/62 ,  G06N3/063 ,  G06N20/00 ,  G06V10/774 ,  G06V10/82 ,  H04L67/00

CPC classification number: H04L67/1097 ,  G06F16/24578 ,  G06F16/438 ,  G06F16/447 ,  G06F18/2148 ,  G06F21/6254 ,  G06N3/063 ,  G06N20/00 ,  G06V10/7747 ,  G06V10/82 ,  H04L67/34

Abstract: Embodiments described herein provide for a non-transitory machine-readable medium storing instructions to cause one or more processors to receive, at a client device, a machine learning model from a server, detect a usage pattern for a content item, store an association between the content item and the detected usage pattern in local data, train the machine learning model using local data for the content item with the detected usage pattern to generate a trained machine learning model, generate an update for the machine learning model, privatize the update for the machine learning model, and transmit the privatized update for the machine learning model to the server.

公开(公告)号：US20220392219A1

公开(公告)日：2022-12-08

申请号：US17658474

申请日：2022-04-08

Applicant: Apple Inc.

Inventor： Michael Chatzidakis ,  Kalu O. Kalu ,  Omid Javidbakht ,  Sowmya Gopalan ,  Eric Circlaeys ,  Rehan Rishi ,  Mayank Yadav

IPC: G06V20/00 ,  G06V20/50 ,  G06F16/906 ,  G06F16/908

Abstract: Devices, methods, and non-transitory program storage devices (NPSDs) are disclosed herein to provide for the privacy-respectful learning of iconic scenes and places, wherein the learning is based on information received from one or more client devices in response to one or more collection criteria specified as part of one or more collection operations launched by a server device. In some embodiments, differential privacy techniques (such as the submission of predetermined amounts of noise-injecting, e.g., randomly-generated, data in conjunction with actual data) are employed by the client devices, such that any insights learned by the server device only relate to “hot spots,” “themes,” or other scenes, objects, and/or topics that are highly popular and captured in the digital assets (DAs) of many users, ensuring there is no way for the server device to learn or glean any insights related to particular users of individual client devices participating in the collection operations.

公开(公告)号：US20210192078A1

公开(公告)日：2021-06-24

申请号：US17129579

申请日：2020-12-21

Applicant: Apple Inc.

Inventor： Stephen Cosman ,  Kalu Onuka Kalu ,  Marcelo Lotif Araujo ,  Michael Chatzidakis ,  Thi Hai Van Do ,  Alexis Hugo Louis Durocher ,  Guillaume Tartavel ,  Sowmya Gopalan ,  Vignesh Jagadeesh ,  Abhishek Bhowmick ,  John Duchi ,  Julien Freudiger ,  Gaurav Kapoor ,  Ryan M. Rogers

IPC: G06F21/62 ,  G06K9/62 ,  H04L29/08 ,  G06N20/00 ,  G06N3/063

Abstract: Embodiments described herein provide for a non-transitory machine-readable medium storing instructions to cause one or more processors to receive, at a client device, a machine learning model from a server, detect a usage pattern for a content item, store an association between the content item and the detected usage pattern in local data, train the machine learning model using local data for the content item with the detected usage pattern to generate a trained machine learning model, generate an update for the machine learning model, privatize the update for the machine learning model, and transmit the privatized update for the machine learning model to the server.

公开(公告)号：US12243308B2

公开(公告)日：2025-03-04

申请号：US17658474

申请日：2022-04-08

Applicant: Apple Inc.

Inventor： Michael Chatzidakis ,  Kalu O. Kalu ,  Omid Javidbakht ,  Sowmya Gopalan ,  Eric Circlaeys ,  Rehan Rishi ,  Mayank Yadav

IPC: G06F16/90 ,  G06F16/906 ,  G06F16/908 ,  G06V20/50

Abstract: Devices, methods, and non-transitory program storage devices (NPSDs) are disclosed herein to provide for the privacy-respectful learning of iconic scenes and places, wherein the learning is based on information received from one or more client devices in response to one or more collection criteria specified as part of one or more collection operations launched by a server device. In some embodiments, differential privacy techniques (such as the submission of predetermined amounts of noise-injecting, e.g., randomly-generated, data in conjunction with actual data) are employed by the client devices, such that any insights learned by the server device only relate to “hot spots,” “themes,” or other scenes, objects, and/or topics that are highly popular and captured in the digital assets (DAs) of many users, ensuring there is no way for the server device to learn or glean any insights related to particular users of individual client devices participating in the collection operations.