Facebook, Inc.

Sung-eok Jeon

Aliasgar Mumtaz Husain

Kevin Russell Wheeler

Techniques to convert multi-party conversations to an editable document are described. In one embodiment, an apparatus may comprise a message queue monitoring component operative to monitor a messaging conversation between two or more messaging participants; and analyze the messaging conversation based on a plurality of domain-specific extractors; a domain selector component operative to determine that a domain detection level for a selected extractor exceeds a threshold; and a document component operative to generate an editable document based on the messaging conversation using the selected extractor and assign access permissions to the editable document automatically to the two or more messaging participants. Other embodiments are described and claimed.

Techniques to convert multi-party conversations to an editable document

Information for an advertising campaign is received that includes one or more advertisements for presentation to one or more users of an online system. A strong interest feature domain is generated for each user of the online system. A set of related items is obtained, wherein each item in the set is associated with an action conducted by the one or more users. It can be determined that a viewing user of the online system has interacted with at least one item in the set of related items, and the generated interest feature domain for the viewing user includes an interest feature domain specified in the advertising campaign. Responsive to the determination, an advertisement is selected for presentation to the viewing user.

Informative advertisements on hobby and strong interests feature space

An online system monitors actions performed by users of the online system in association with being presented with content items that are associated with various categories. A histogram may be generated to describe a pattern of an action previously performed by a user in response to being presented with content items associated with a category. The online system trains a model to predict a likelihood that the user will perform the action in response to being presented with a content item associated with the category based on the pattern of the action previously performed by the user and information describing one or more recent performances of the action by the user with a content item associated with the category. The predicted likelihood may be included in a content selection process that selects one or more content items for presentation to the user.

Selecting content for presentation to an online system user based on categories associated with content items

In one embodiment, a method includes receiving from a first user a text query inputted by the first user using an I/O device associated with the client system, the I/O device including a first keyboard layout. The method also includes identifying one or more physical-proximity phrases corresponding to the text query based at least in part on the first keyboard layout and calculating a confidence score for each of the identified physical-proximity phrases based at least in part on a search history of the first user. The method also includes identifying one or more objects, where each identified object matches at least one of the identified physical-proximity phrases having a confidence score greater than a threshold confidence score and sending to the first user one or more search results corresponding to one or more of the identified objects, respectively.

Keyboard-based corrections for search queries on online social networks

An online system identifies a group of users of the online system based on connections between the users of the group maintained by the online system and distances between the users during a particular time interval. The online system predicts a purpose of the group and one or more common interests shared by users of the group. The predicted purpose and common interest(s) are mapped to content items eligible to be presented to one or more users of the group. A confidence score and/or bid price is computed for each content item to which the purpose and common interest(s) of the group are mapped. The online system selects one or more content items for presentation to one or more users of the group based on the confidence scores/bid prices.

Selecting content for presentation to a group of online system users based on temporal, spatial, interest, and purpose information of the group

An online system receives content items including targeting criteria and associated with objects presented via video content. When the online system presents video content to a user, the online system detects objects included in the video content. If the online system detects an object in the video content matching an object associated with a content item, the online system determines whether characteristics of the user satisfy at least a threshold amount of targeting criteria included in the content item. In response to characteristics of the user satisfying the threshold amount of targeting criteria included in the content item, the online system includes the content item in one or more selection processes. If the selection processes select the content item, the online system presents the content item in conjunction with the video content.

Selecting content items for presentation in conjunction with video content presented to a user based on objects identified in the video content

A method trains a model for providing content items to users of a social networking system. The system generates profile vectors based on user profile information such as demographic data and personal data. The system logs actions performed by users on the social networking system and generates behavior vectors based on the logged actions. The profile vectors and behavior vectors are each associated with a user attribute, e.g., the age or gender of a user. The system generates a difference vector based on a profile vector and a behavior vector. The difference vector is then used as a feature to train the model using machine learning techniques. The trained model may select content items that a target user is most likely to be interested in and interact with.

Statistical feature engineering of user attributes

An online system automatically and dynamically determines an audience for content by clustering users across various dimensions, and refining targeting criteria for the content. The online system receives content and initial targeting criteria from a content provider. The content is provided to a group of users that meet the initial targeting criteria. The system collects content response data from the group of users that were provided the content, including user responses to the content and dynamic data relating to time and location of the user responses. The content response data is further integrated with user characteristics, content presentation data, and social response data to generate integrated user-content data of the content. Clusters of users are generated based on features of the integrated user-content data, and refined targeting criteria are identified based on the generated clusters that can then be used for more accurate targeting of the content to users.

Growing community of inventors

Bellevue, WA, United States of America

Sung-eok Jeon