Abstract: In an example embodiment, a single machine learned model that allows for ranking of entities across all of the different combinations of node types and edge types is provided. The solution calibrates the scores from Edge-FPR models to a single scale. Additionally, the solution may utilize a per-edge type multiplicative factor dictated by the true importance of an edge type, which is learned through a counterfactual experimentation process. The solution may additionally optimize on a single, common downstream metric, specifically downstream interactions that can be compared against each other across all combinations of node types and edge types.

Abstract: In an example embodiment, a single machine learned model that allows for ranking of entities across all of the different combinations of node types and edge types is provided. The solution calibrates the scores from Edge-FPR models to a single scale. Additionally, the solution may utilize a per-edge type multiplicative factor dictated by the true importance of an edge type, which is learned through a counterfactual experimentation process. The solution may additionally optimize on a single, common downstream metric, specifically downstream interactions that can be compared against each other across all combinations of node types and edge types.

Abstract: Methods, systems, and computer programs are presented for capturing teammate-connection information and leveraging this information to improve services of a social networking service. The teammates feature allows users to identify a subset of their connections as teammate connections, which are connections that work directly with the user as a manager, a peer (person reporting to the same manager as the user), or a direct report. The teammate-connection information is used for improving services and notifications, such as new-connection suggestions, construction of the user feed, group conversations and messaging, notification prioritization, etc. Unlike existing connection patterns, the teammate-connections information is structured data. The structured data provides a better understanding of the relationships among the users and allows optimization of the user experience. In one aspect, a method is presented to generate recommendations of possible teammates for users of the social networking service.