Abstract: Disclosed here is a system that can obtain attributes of an advertisement, where an attribute has a continuous value, and a range of acceptable values is uncertain. The system can create a file including contents that when provided to a predetermined function produce a value of the attribute. Based on the file, the system can generate values corresponding to the attributes. Based on the generated values, the system can create the advertisement. The system can obtain a response data to the created advertisement and can fit a multidimensional function to the attributes and the user response data. Based on the multidimensional function, the system can determine next values and next ranges, where the next values and the next ranges indicate an improvement in the response data.

Abstract: An online system generates predicted outcomes for a content distribution program that distributes content to users of the online system, the predicted outcome indicating a likelihood for the occurrence of an outcome of a content presentation. The online system transmits the one or more predicted outcomes to the third-party system, and receives prediction improvement data from the third-party system, the prediction improvement data indicating an adjustment to errors in the predicted outcomes based on a prediction by the third-party system. The online system updates the properties of a content distribution program based on the prediction improvement data, the updated content distribution program causing the online system to generate new predicted outcomes based on the prediction improvement data in content presentation opportunities. The online system also transmits content to users of the online system based on the updated content distribution program.

Abstract: An online system accesses a list of features used as input into a predictor to predict a performance metric for content presented to users. The online system computes importance scores for one or more of the features. A ranked list of categories is created, with each category having one or more sub-categories. For each feature having a computed importance score, the online system assigns, for each attribute in the ranked list of attributes for that feature, the feature to a sub-category in one of the categories in the ranked list of categories that has the same rank as the attribute in the ranked list of attributes for the feature, where the sub-category is associated with a label that corresponds with the attribute. For each sub-category in each category, a cumulative score is computed for the sub-category based on the importance scores of the features of that sub-category.

Abstract: An online system ranks content eligible for presentation to an online system user based on a prediction made by a general model or a specific model indicating a likelihood that the user will interact with a content item, in which the specific model has a higher latency than the general model. The online system determines which prediction to use for the ranking by balancing the benefit of a more accurate prediction made by the specific model against the higher latency of the specific model. The online system outputs the predicted likelihood from one of the models based on the determination, ranks content items eligible for presentation to the user based on the output, and selects content item(s) for presentation to the user based on the ranking. The online system may log the predicted likelihoods from both models, the outputted predicted likelihood, and information describing the performance of the content item.

Abstract: A system receives a first plurality of impressions associated with a first set of features. Labels for the first plurality is generated based on the first set of features. A machine learning model is trained based on the first set of features and the labels. A second plurality of impressions associated with a second set of features is received. A first estimated probability measuring conversion likelihood when impression delivery occurs is generated based on applying the second plurality and the second set of features to the model. A plurality of holdout impressions associated with a third set of features is identified. A second estimated probability measuring conversion likelihood when impression delivery is withheld is generated based on applying the plurality of holdout impressions and the third set of features to the model. Increased valued (e.g., lift) is estimated based on subtracting the second estimated probability from the first estimated probability.

Abstract: Disclosed here is a system that can obtain attributes of an advertisement, where an attribute has a continuous value, and a range of acceptable values is uncertain. The system can create a file including contents that when provided to a predetermined function produce a value of the attribute. Based on the file, the system can generate values corresponding to the attributes. Based on the generated values, the system can create the advertisement. The system can obtain a response data to the created advertisement and can fit a multidimensional function to the attributes and the user response data. Based on the multidimensional function, the system can determine next values and next ranges, where the next values and the next ranges indicate an improvement in the response data.

Abstract: Methods and systems provide for a unified presentation of cross-platform content to a user visiting a platform. In one embodiment, the system connects a client device associated with a user to a first content platform; receives a request from the client device to present content to the user at the first content platform; receives content associated with one or more additional content platforms; determines a subset of the content to present to the user; standardizes the subset of the content in a format to be used at the first content platform; presents the subset of the content to the user at the first content platform; processes a set of unified cross-platform metrics for user events related to the user interacting with the subset of the content at the first content platform; and provides a report of the set of unified cross-platform metrics for the user events.

Abstract: Disclosed herein is a system and method to determine whether to place an advertisement to a user requesting an address from the user. The system can iteratively determine multiple advertisement metrics of multiple advertisements to obtain multiple metrics. An advertisement metric among the multiple advertising metrics can indicate the value of placing the advertisement to the user. The system can rank multiple advertisements based on the multiple advertisement metrics and present a predetermined percentage of top-ranking advertisements among the multiple advertisements.

Abstract: An online system accesses a model attribute store, which stores configuration information and model performance scores for a plurality of models, each model used to predict performance metrics regarding content from a third party system presented to users of the online system. The online system trains a meta-model classifier using the models in the model attribute store, the meta-model classifier trained to predict, for a candidate model, a predicted model performance score of that candidate model. The online system also generates a plurality of candidate models for input to the meta-model classifier, each of the plurality of candidate models including a distinct set of configuration information. The predicted model performance scores for a selected candidate model in the plurality of candidate models is computed using the meta-model classifier, and the online system transmits a report to the third party system indicating predicted model performance score for the selected candidate model.

Abstract: An online system generates predicted outcomes for a content distribution program that distributes content to users of the online system, the predicted outcome indicating a likelihood for the occurrence of an outcome of a content presentation. The online system transmits the one or more predicted outcomes to the third-party system, and receives prediction improvement data from the third-party system, the prediction improvement data indicating an adjustment to errors in the predicted outcomes based on a prediction by the third-party system. The online system updates the properties of a content distribution program based on the prediction improvement data, the updated content distribution program causing the online system to generate new predicted outcomes based on the prediction improvement data in content presentation opportunities. The online system also transmits content to users of the online system based on the updated content distribution program.

Abstract: An online system, such as a social networking system, generates shared models for one or more clusters of categories. A shared model for a cluster is common to the categories assigned to the cluster. In this manner, the shared models are specific to the group of categories (e.g., selected content providers) in each cluster while requiring a reasonable computational complexity for the online system. The categories are clustered based on the performance of a model specific to a category on data for other categories.

Abstract: An online system receives explicit user data and explicit event data, and implicit user data and implicit event data from a third party system. The online system generates an implicit users/implicit events data feature, an explicit users/explicit events data feature, and an explicit users/implicit events data feature. The online system generates a prediction of the counterfactual rate based on the implicit users/implicit events data feature, the explicit users/explicit events data feature, and the explicit users/explicit events data feature, the counterfactual rate indicating the likelihood that target users matching certain characteristics caused an event to occur when the target are not been presented with content by the online system, the content configured to induce users to cause the event to occur. A combined prediction rate is presented to the third party system based on the counterfactual rate.

Abstract: An online system that identifies allocations of both organic and promoted content on a given page. The allocations of page content are compared against one another and configured to prioritize for overall utility based on objective factors that quantify a page “look and feel” as measured by machine learning models. The page allocations are operated on an automatic and continuous basis for each user viewing the page. In some embodiments, the page content allocations are based on individual viewing users stored characteristics.

Abstract: Disclosed here is a system that can obtain attributes of an advertisement, where an attribute has a continuous value, and a range of acceptable values is uncertain. The system can create a file including contents that when provided to a predetermined function produce a value of the attribute. Based on the file, the system can generate values corresponding to the attributes. Based on the generated values, the system can create the advertisement. The system can obtain a response data to the created advertisement and can fit a multidimensional function to the attributes and the user response data. Based on the multidimensional function, the system can determine next values and next ranges, where the next values and the next ranges indicate an improvement in the response data.

Abstract: An online system generates content feature entries, with each content feature entry describing a content item from a third party system. The online system generates user feature entries, each user feature entry describing a user. The online system generates a combination score for a target user and a selected content item by computing a combination of the content feature entries associated with the selected content item and the user feature entries associated with the target user using a combining function. The combination score indicates an estimated increase in value for the third party system when the target user is presented with the selected content item. The online system selects content items to transmit to a client device of a target user of the online system for presentation to the target user based on the combination score for the content items and the target user.

Abstract: An online system uses multiple machine learning models to select content for providing to a user of the online system. Specifically, the online system trains a general model that intakes a first set of features and outputs predictions at a general level. The online system further trains a residual model that intakes a second set of features. The residual model predicts a residual (e.g., an error) of the predictions outputted by the general model. Therefore, the predicted residual from the residual model is combined with the prediction from the general model in order to correct for the over-generality of the general model. The online system may use the combined prediction to send content to users.

Abstract: An online system generates predicted outcomes for a content distribution program that distributes content to users of the online system, the predicted outcome indicating a likelihood for the occurrence of an outcome of a content presentation. The online system transmits the one or more predicted outcomes to the third party system, and receives prediction improvement data from the third party system, the prediction improvement data indicating an adjustment to errors in the predicted outcomes based on a prediction by the third party system. The online system updates the properties of a content distribution program based on the prediction improvement data, the updated content distribution program causing the online system to generate new predicted outcomes based on the prediction improvement data in content presentation opportunities. The online system also transmits content to users of the online system based on the updated content distribution program.

Abstract: An online system, such as a social networking system, generates shared models for one or more clusters of categories. A shared model for a cluster is common to the categories assigned to the cluster. In this manner, the shared models are specific to the group of categories (e.g., selected content providers) in each cluster while requiring a reasonable computational complexity for the online system. The categories are clustered based on the performance of a model specific to a category on data for other categories.

Abstract: An online system optimizes selection content items for a user based on total value of presenting a selected content item, rather than discrete actions with the content item. To account for the total value of presenting a content item, the online system receives information from a third party system associated with the content item identifying actions by users captured by the third party system and values associated with the identified actions. The online system matches the identified actions with presentations of the content item to various users by identifying users of the online system corresponding to information identifying users received from the third party system and retrieves information describing presentation of content items to the information identifying presentation of content items. Based on historical actions and presentations of a content item, the online system obtains a model determining value of presenting a content item for use in selecting content.

Abstract: An online system determines how presenting an awareness campaign to a user will affect the user's likelihood of converting to a related direct response campaign. For the user, the online system creates a benchmark exposure profile representing the user's exposure history before the awareness campaign. Similarly, the online system determines the user's simulated exposure profile, which represents the user's brand exposure history after having been exposed to the awareness campaign. A response prediction for the direct response campaign is determined for the benchmark exposure profile and the simulated exposure profile. The online system estimates the difference between the response prediction and the simulated response prediction to determine a delivery control value of presenting the awareness campaign to a user. The delivery control value is used to determine an effective impression value for the awareness campaign and conversion value for the related direct response campaign.