Abstract: Embodiments of the disclosed technologies receive first event data associated with a first party application, receive second event data representing a click, in the first party application, on a link to a third party application, receive third event data from the third party application, convert the third event data to a label, map a compressed format of the labeled third event data to the first event data and the second event data to create multi-party attribution data, group multiple instances of the multi-party attribution data into a batch, add noise to the compressed format of the labeled third event data in the batch, and send the noisy batch to a second computing device. A debiasing algorithm can be applied to the noisy batch. The debiased noisy batch can be used to train at least one machine learning model.

Abstract: Techniques for controlling item frequency using machine learning are provides. In one technique, two prediction models are trained: one based on interaction history of multiple content items by multiple entities and the other based on predicted interaction rates and an impression count for each of multiple content items. In response to a request, a particular entity associated with the request is identified and multiple candidate content items are identified. For each identified candidate content item, the first prediction model is used to determine a predicted interaction rate, an impression count of the candidate content item is determined with respect to the particular entity, the second prediction model is used to generate an adjustment based on the impression count, and an adjusted entity interaction rate is generated based on the predicted interaction rate and the adjustment. A particular candidate content item is selected based on the generated adjusted entity interaction rates.

Abstract: Techniques for controlling resource usage in a computing environment are provided. In one technique, a target resource usage for a particular point in time is determined for a content delivery campaign. Determining, for the content delivery campaign, a current resource usage for the particular point in time. Also, a bandwidth associated with the target resource usage at the particular point in time is determined. Based on a difference between the current resource usage and one or more boundaries of the bandwidth, a throttling factor is calculated. Based on the throttling factor, a probability of the content delivery campaign participating in a content item selection event is determined.

Abstract: Techniques for using attention events for audience expansion are provided. In one technique, first interaction data that indicates multiple interactions by the first entity with multiple content items is stored. The interactions includes an interaction that is based on an amount of time that content within one of the content items was presented to the first entity. Based on the first interaction data, similarity data that identifies one or more content delivery campaigns that are similar to a particular content delivery campaign is generated. Second interaction data that indicates interaction(s) by a second entity with content item(s) is stored. Based on the second interaction data and the similarity data, association data that associates the second entity with the particular content delivery campaign is stored. The association data may be used to identify the particular campaign in response to receiving a content request from a computing device of the second entity.

Abstract: Techniques for controlling item frequency using machine learning are provides. In one technique, two prediction models are trained: one based on interaction history of multiple content items by multiple entities and the other based on predicted interaction rates and an impression count for each of multiple content items. In response to a request, a particular entity associated with the request is identified and multiple candidate content items are identified. For each identified candidate content item, the first prediction model is used to determine a predicted interaction rate, an impression count of the candidate content item is determined with respect to the particular entity, the second prediction model is used to generate an adjustment based on the impression count, and an adjusted entity interaction rate is generated based on the predicted interaction rate and the adjustment. A particular candidate content item is selected based on the generated adjusted entity interaction rates.

Abstract: Techniques for accounting for position-specific differences in user interaction while conducting content item selection events are provided. In one technique, a position-specific factor is determined. The position-specific factor may be based on a ratio of an observed interaction and a predicted interaction. Different positions in a content item feed or on a web page may be associated with different position-specific factors. Eventually, multiple content items are identified for presentation on a screen of a computing device. The content items include a first content item for which a predicted interaction rate is calculated and a second content item for which no predicted interaction rate is calculated. An order of the content items is determined based on the position-specific factor. For example, the predicted interaction rate of the first content item is modified based on the position-specific factor. The content items are presented on the screen based on the order.

Abstract: Techniques for accounting for position-specific differences in user interaction while conducting content item selection events are provided. In one technique, a position-specific factor is determined. The position-specific factor may be based on a ratio of an observed interaction and a predicted interaction. Different positions in a content item feed or on a web page may be associated with different position-specific factors. Eventually, multiple content items are identified for presentation on a screen of a computing device. The content items include a first content item for which a predicted interaction rate is calculated and a second content item for which no predicted interaction rate is calculated. An order of the content items is determined based on the position-specific factor. For example, the predicted interaction rate of the first content item is modified based on the position-specific factor. The content items are presented on the screen based on the order.

Abstract: Techniques are provided for automatically adjusting a resource reduction amount based on resource availability and other factors. The following are determined for a content delivery campaign: a winning distribution for a target audience of the content delivery campaign, a through rate distribution of the content delivery campaign, a resource allocation of the content delivery campaign, and an estimated number of content item selection events in which the content delivery campaign will participate in a future time period. Based on these values, a resource reduction amount is determined. An example of a resource reduction amount is an effective cost per impression. The resource reduction amount is used in one or more subsequent content item selection events in which the content delivery campaign participates.

Abstract: Techniques for controlling resource usage in a computing environment are provided. In one technique, a target resource usage for a particular point in time is determined for a content delivery campaign. Determining, for the content delivery campaign, a current resource usage for the particular point in time. Also, a bandwidth associated with the target resource usage at the particular point in time is determined. Based on a difference between the current resource usage and one or more boundaries of the bandwidth, a throttling factor is calculated. Based on the throttling factor, a probability of the content delivery campaign participating in a content item selection event is determined.

Abstract: The present application relates to systems and computer-implemented methods for calculating a suggested reserve price associated with an opportunity to realize an online advertisement. In some implementations, data associated with historical online advertisement auctions are clustered based on a reference opportunity for realizing an online advertisement, wherein each historical online advertisement auction of the cluster is associated with a first bid price of higher value and a second bid price of lower value; a dominance relationship between a distribution of the first bid prices of the cluster and a distribution of the second bid prices of the cluster is determined; a reserve price associated with the cluster based on the dominance relationship is calculated; and the reserve price is stored as a suggested reserve price to realize an online advertisement, wherein the suggested reserve price is associated with the reference opportunity that is associated with the cluster.

Abstract: The present application relates to systems and computer-implemented methods for calculating a suggested market variable associated with an auction of an online advertisement realization opportunity. In some implementations, an optimization procedure may be operated, wherein the optimization procedure may comprise sending an initial market variable associated with one or more auctions to bidding agents; receiving from each of the bidding agents a response market variable associated with the initial market variable; determining according to an auction bidding rule a winning market variable from the response market variables; and substituting the initial market variable with the wining market variable. The optimization procedure may be operated repeatedly until the winning market variable stabilizes. Then the stabilized market variable may be sent to an advertiser as the suggested market variable.

Abstract: The present application relates to systems and computer-implemented methods for determining a future bidding strategy regarding auctions associated with realization of an online advertisement. In some implementations, a server may be used to determine an optimized bid pacing parameter for an online advertisement based on a current bid pacing parameter associated with the online advertisement and an online advertisement supply information for a time period so that based on the optimized bid pacing, realization of the online advertisement over the time period substantially confirms to a spending strategy associated with the online advertisement during the time period. As such, the systems and computer-implemented methods may allow advertisers to control their spending strategy accurately during an advertising campaign.

Abstract: The present application relates to systems and computer-implemented methods for calculating a suggested reserve price associated with an opportunity to realize an online advertisement. In some implementations, a database of historical online advertisement auctions is established; historical online advertisement auctions from the database of historical online advertisement auctions that are associated with a feature are clustered to form a cluster; a reserve price associated with the cluster of historical online advertisement auctions is calculated to generate a desired revenue; and the reserve price is stored as a suggested reserve price for the opportunity to realize the online advertisement that is associated with the feature.

Abstract: The present disclosure generally relates to ad auction optimization. In some examples, methods, systems, and computer programs for ad auction optimization using machine learning algorithms to estimate a likelihood that a consumer will purchase an advertised product and balance long term and short term goals to determine modeled data for a keyword in an auction are described.