Facebook, Inc.

Rajiv Krishnamurthy

Xiaojun Liang

Lin Qiao

Hongzhong Jia

Joshua David Metzler

A dataset management system ('system') reduces the amount of data to be stored for future analyses. The system determines a sampling rate of the data based on a required level of accuracy, and samples the data at the determined sampling rate. Initially, all data transactions ('full dataset') and the sampled data (&#x201c;sampled dataset&#x201d;) are logged and stored. Based upon a trigger condition, e.g., after a specified period, the full dataset and the sampled dataset are analyzed separately and the analysis results are compared. If the comparison is sufficiently similar (i.e., the sampling produces a sufficiently accurate set of data or a variance between the analysis results of the datasets is within a specified threshold), the system discontinues full data logging and stores only the sampled dataset. Further, the full dataset is deleted. The sampling thus reduces the required data volume significantly, thereby minimizing consumption of the storage space.

Reducing data storage requirements

A client device or an online system determines a position in an interface presented by a client device for presenting content associated with an application installed on the client device based in part on a bid amount associated with the application. Scores are determined for the application and other applications installed on the client device based on an expected amount of user interaction with each application and bid amounts associated with one or more of the applications. Based on a score associated with an application, a position in the interface for presenting content associated with the application is determined. If the determined position satisfies a position specified by the bid amount and content associated with the application is presented in the determined position for at least a threshold amount of time, an online system charges a third party system associated with the application an amount.

Bidding for position of installed application in an interface of a mobile device

The disclosure is related to encoding a message or portions of the message to be exchanged between a pair of users. Messages can be exchanged between a pair of users in multiple ways, e.g., electronic mail, short messaging service (SMS), messenger applications. The computing resource, e.g., network bandwidth consumed in transmission of the messages, can be minimized by encoding/compressing at least portions of a message prior to transmission. A server analyzes messages exchanged between a pair of users to identify message portions, e.g., a word or a phrase, that are commonly exchanged and generates a set of codes for those message portions. The pair of users can use the set of codes for encoding and/or decoding the messages exchanged between them. The codes are generated using a specified coding technique, e.g., Huffman coding.

Encoding portions of a message

Various embodiments relating to a social cache replacement policy are described. The techniques of the present invention disclosed utilize social network properties to guide a cache replacement policy executed by a social networking platform system. In one embodiment, a method is provided for determining a queue location to cache a data item based on a popularity score computed from social network properties. In one embodiment, a method is provided for computing the popularity score by incorporating a user's social network properties and the user's friends' social network properties. In embodiments, the popularity score may be computed using a plurality of social network properties, which may include social network properties associated with (i) the user, (ii) the consumer(s), and/or (iii) the data item(s). In embodiments, a plurality of popularity scores are maintained in a user-score database, where the plurality of popularity scores are periodically updated using historical data.

Growing community of inventors

Sunnyvale, CA, United States of America

Rajiv Krishnamurthy