Facebook, Inc.

Ashish Kumar Yadav

Bradley Ray Green

Jinyi Yao

James Wah Hou Wong

Brian Piepgrass

Tony Hsien-yu Liu

Julia Lee

Daniel Dinu

Deeptanshu Verma

Neal Suresh Vora

Lingjuan Peng

Niting Qi

Andre Nader

Systems, methods, and non-transitory computer readable media can determine one or more user-related metrics relating to each page of a plurality of pages associated with an administrator based on a first machine learning model. One or more recommendations relating to each page of the plurality of pages can be determined based on a second machine learning model. One or more pages of the plurality of pages for which to display cards including page updates in a feed of the administrator can be determined, based on the determined user-related metrics and the determined recommendations.

Systems and methods for providing feed page updates in a social networking system

The social networking system monitors implicit interactions between a user and objects of the social networking system with which the user has not established a connection. Based on the implicit interactions between the user and an object, the social networking system identifies a soft connection between the user and the object. The social networking system may then identify soft connections to include in a candidate list of soft connections to recommend to the user. The social networking system may also extract signals from the set of candidate list of soft connections, and may use the extracted signals to rank the soft connections in the list of candidate soft connections. The social networking system may then recommend soft connections to the user based on the rank associated with the soft connections in the candidate list of soft connections.

Recommending objects to a user of a social networking system based on implicit interactions between the user and the recommended objects

A method for making entity recommendations to users of a social networking system (SNS) by maximizing a predicted benefit to the SNS through objective value models. The method includes generating a plurality of candidate entities eligible for recommendation, and determining a weight for each candidate entity based upon an output score generated by an objective value model. The objective value model utilizes objectively measurable input values describing the candidate entities and generates larger output scores for those candidate entities that would create larger predicted increases in an objectively measurable amount of benefit for the SNS created by a potential connection being created between the user and the respective candidate entity. One or more entities are selected based upon the determined weights, and an entity recommendation user interface is presented to the user that includes one or more recommendation modules identifying the selected one or more entities.

Objective value models for entity recommendation

Growing community of inventors

Mountain View, CA, United States of America

Ashish Kumar Yadav