Flipboard, Inc.

Robert Griesmeyer

A digital magazine server scores digital magazines based on how related the digital magazines are to each of a set of topics in a topic vector space, which allows ranking for each topic the magazines that are most closely related to the topic. The digital magazine server generates a magazine vector for the magazine in the topic vector space by aggregating extracted topics for the magazine and compares the magazine vector to each of the topics to determine a magazine-topic relevance score that indicates the relevance of the digital magazine to each topic in the topic vector space. The digital magazine server then ranks each of the digital magazines by their magazine-topic relevance scores for a particular topic, e.g., a trending topic from a user request, and selects a top number of magazines for the trending topic for the user.

Digital magazine recommendations by topic

A system generates and displays to a user of a digital magazine a list of candidate topics. The system first receives a package name list from a user device of the user, and the package name list indicates computer applications installed on the user device. The system retrieves a package name-topic matrix generated based on information associated with a set of existing users of the system. The system compares package names included in the received package name list and package names included in the package name-topic matrix, and identifies known topics in the package name-topic matrix based on the comparison. The system generates a list of candidate topics for the user based on the identified known topics, and provides the list for display to the user. In response to candidate topics selected by the user, the system provides to the user content relevant to the selected candidate topics.

Candidate topic picker for digital magazine

The disclosure relates (a) a method and computer program product for training a content classifier and (b) a method and computer program product for using the trained content classifier to determine compliance of content items with a content policy of an online system. A content classifier is trained using two training sets, one containing NSFW content items and the other containing SFW content items. Content signals are extracted from each content item and used by the classifier to output a decision, which is compared against its known classification. Parameters used in the classifier are adjusted iteratively to improve accuracy of classification. The trained classifier is then used to classify content items with unknown classifications. Appropriate action is taken for each content item responsive to its classification. In alternative embodiments, multiple classifiers are implemented as part of a two-tier classification system, with text and image content classified separately.

Growing community of inventors

San Francisco, CA, United States of America

Robert Griesmeyer