Low entropy browsing history for content quasi-personalization
AssigneeGoogle Inc.
InventorsGang Wang, Marcel M.M. Yung
Life of this patent
Estimating the term…
Abstract
The present disclosure provides systems and methods for content quasi-personalization or anonymized content retrieval via aggregated browsing history of a large plurality of devices, such as millions or billions of devices. A sparse matrix may be constructed from the aggregated browsing history, and dimensionally reduced, reducing entropy and providing anonymity for individual devices. Relevant content may be selected via quasi-personalized clusters representing similar browsing histories, without exposing individual device details to content providers.
Claims21
(canceled)
A method for anonymized content retrieval, comprising:
- generating, by an application executing at a client device, a set of feature vectors representing online content accessed by the client device;
- reducing, by the application executing at the client device, a number of dimensions of the set of feature vectors to obtain a reduced set of feature vectors;
- generating, by the application executing at the client device, a class identifier based on the reduced set of feature vectors;
- generating, by the application executing at the client device, a content request including the class identifier;
- transmitting, by the application executing at the client device, the content request including the class identifier to a content server;
- and
- receiving, by the application executing at the client device and from the content server, an item of content selected according to the class identifier.
A system for anonymized content retrieval, comprising:
- a client device executing an application;
- a memory storing instructions that cause the client device to perform operations comprising:
- generating a set of feature vectors representing online content accessed by the client device;
- reducing a number of dimensions of the set of feature vectors to obtain a reduced set of feature vectors;
- generating a class identifier based on the reduced set of feature vectors;
- generating, by the application executing at the client device, a content request including the class identifier;
- transmitting the content request including the class identifier to a content server;
- and
- receiving, from the content server, an item of content selected according to the class identifier.
A non-transitory computer readable medium storing instructions that, upon execution, cause one or more data processing apparatus to perform operations comprising:
- generating a set of feature vectors representing online content accessed by the client device;
- reducing a number of dimensions of the set of feature vectors to obtain a reduced set of feature vectors;
- generating a class identifier based on the reduced set of feature vectors;
- generating, by the application executing at the client device, a content request including the class identifier;
- transmitting the content request including the class identifier to a content server;
- and
- receiving, from the content server, an item of content selected according to the class identifier.
Dependencies are inferred from the claim text.