Abstract: Machine learning techniques for multi-objective content item selection are provided. In one technique, resource allocation data is stored that indicates, for each campaign of multiple campaigns, a resource allocation amount that is assigned by a central authority. In response to receiving the content request, a subset of the campaigns is identified based on targeting criteria. Multiple scores are generated, each score reflecting a likelihood that a content item of the corresponding campaign will be selected. Based on the scores, a particular campaign from the subset is selected and the corresponding content item transmitted over a computer network to be displayed on a computing device. A resource allocation amount that is associated with the particular campaign is identified. A resource reduction amount associated with displaying the content item of the particular campaign is determined. The particular resource allocation is reduced based on the resource reduction amount.

Abstract: Techniques are described herein for determining whether to provide an advertisement to a user of a social network. The determination is based on a click probability and a social network value for the user. The click probability indicates a likelihood of the user to select the advertisement if provided to the user via the social network. The social network value is based on a subscription probability of the user and further based on subscription probabilities of other users in the social network that are included in an affinity set of the user. Each subscription probability indicates a likelihood of a respective user to subscribe to a paid service with respect to the social network.

Abstract: According to some example embodiments, a method includes calculating learning values associated with a plurality of listings, at least one of said learning values associated with one of said listings representing a value based, at least in part, on a probability distribution of selections of said listing. The method further includes applying said learning values to ranking scores associated with said listings to provide an updated ranking, and electronically auctioning advertising inventory to purchasers associated with said listings based, at least in part, on said updated ranking.

Abstract: Software for initialized explore-exploit creates a plurality of probability distributions. Each of these probability distributions is generated by inputting a quantitative description of one or more features associated with an image into a regression model that outputs a probability distribution for a measure of engagingness for the image. Each of the images is conceptually related to the other images. The software uses the plurality of probability distributions to initialize a multi-armed bandit model that outputs a serving scheme for each of the images. Then the software serves a plurality of the images on a web page displaying search results, based at least in part on the serving scheme.

Abstract: A set of item-item affinities for a plurality of items is determined based on collaborative-filtering techniques. A set of an item's nearest neighbor items based on the set of item-item affinities is determined. A set of user feature-item affinities for the plurality of items and a set of user features is determined based on least squared regression. A set of a user feature's nearest neighbor items is determined based in part on the set of user feature-item affinities. Compatible affinity weights for nearest neighbor items of each item and each user feature are determined. Based on user features of a user and items a user has consumed, a set of nearest neighbor items are identified as a set of candidate items, and affinity scores of candidate items are determined. Based on the affinity scores, a candidate item from the set of candidate items is recommended to the user.

Abstract: Methods and systems are provided that may be used to determine a probability of whether a visitor to a web document is likely to click on a web advertisement. An exemplary method may include detecting one or more features in a web document. One or more expert statistical models to which the web document belongs may be determined and associated weightings may be determined based, at least in part, on the one or more features detected. A click-through-rate probability for a web advertisement to be placed on the web document may be estimated based on the one or more expert statistical models.

Abstract: Techniques are described herein for determining whether to provide an advertisement to a user of a social network. The determination is based on a click probability and a social network value for the user. The click probability indicates a likelihood of the user to select the advertisement if provided to the user via the social network. The social network value is based on a subscription probability of the user and further based on subscription probabilities of other users in the social network that are included in an affinity set of the user. Each subscription probability indicates a likelihood of a respective user to subscribe to a paid service with respect to the social network.

Abstract: Content items are selected to be displayed on a portal page in such a way as to maximize a performance metric such as click-through rate. Problems relating to content selection are addressed, such as changing content pool, variable performance metric, and delay in receiving feedback on an item once the item has been displayed to a user. An adaptation of priority-based schemes for the multi-armed bandit problem, are used to project future trends of data. The adaptation introduces experiments concerning a future time period into the calculation, which increases the set of data on which to solve the multi-armed bandit problem. Also, a Bayesian explore/exploit method is formulated as an optimization problem that addresses all of the issues of content item selection for a portal page. This optimization problem is modified by Lagrange relaxation and normal approximation, which allow computation of the optimization problem in real time.

Abstract: Software for initialized explore-exploit creates a plurality of probability distributions. Each of these probability distributions is generated by inputting a quantitative description of one or more features associated with an image into a regression model that outputs a probability distribution for a measure of engagingness for the image. Each of the images is conceptually related to the other images. The software uses the plurality of probability distributions to initialize a multi-armed bandit model that outputs a serving scheme for each of the images. Then the software serves a plurality of the images on a web page displaying search results, based at least in part on the serving scheme.

Abstract: Techniques are described herein for facilitating the consumption of user-generated comments by determining which comments will be of most interest to each individual user. Once the comments that will be of most interest to a particular user are determined, the user-generated comments are presented to that user in a manner that reflects that user's predicted interest. A variety of factors may be used to predict, automatically, the interest each individual user would have in each user-generated comment. For example, interest predictions for a user may be based on the user's prior rating of comments, various types of profile and/or demographic information about the user, the user's social network connections, the authors of the comments, the author of the target subject matter, the user's propensity to comment, etc.

Abstract: Techniques are described herein for determining whether to provide an advertisement to a user of a social network. The determination is based on a click probability and a social network value for the user. The click probability indicates a likelihood of the user to select the advertisement if provided to the user via the social network. The social network value is based on a subscription probability of the user and further based on subscription probabilities of other users in the social network that are included in an affinity set of the user. Each subscription probability indicates a likelihood of a respective user to subscribe to a paid service with respect to the social network.

Abstract: Methods and systems are provided for click through rate prediction and advertisement selection in online advertising. Methods are provided in which output information from a feature-based machine learning model is utilized. The output information includes predicted click through rate information. The output information is used to form a matrix. The matrix is modeled using a latent variable model. Machine learning techniques can be used in determining values for unfilled cells of one or more model matrices. The latent variable model can be used in determining predicted click through rate information, and in advertisement selection in connection with serving opportunities.

Abstract: Content items are selected to be displayed on a portal page in such a way as to maximize a performance metric such as click-through rate. Problems relating to content selection are addressed, such as changing content pool, variable performance metric, and delay in receiving feedback on an item once the item has been displayed to a user. An adaptation of priority-based schemes for the multi-armed bandit problem, are used to project future trends of data. The adaptation introduces experiments concerning a future time period into the calculation, which increases the set of data on which to solve the multi-armed bandit problem. Also, a Bayesian explore/exploit method is formulated as an optimization problem that addresses all of the issues of content item selection for a portal page. This optimization problem is modified by Lagrange relaxation and normal approximation, which allow computation of the optimization problem in real time.

Abstract: An improved system and method for forecasting an inventory of online advertisement impressions for targeting profiles of attributes is provided. An index of advertisement impressions on display advertising properties may be built for a targeting profile of attributes from forecasted impression pools. Impression pools of advertisements sharing the same attributes and trend forecast data for web pages and advertisement placements on the web pages may be integrated to generate the forecasted impression pools. An index of several index tables may be generated from forecasted impression pools. A query may be submitted to obtain an inventory forecast of advertisement impressions for targeting profiles of attributes and the index may be searched to match forecasted impression pools for the targeted profile of attributes. Then the inventory forecast of advertisement impressions on display advertising properties may be returned as query results for the targeting profile of attributes.

Abstract: A set of item-item affinities for a plurality of items is determined based on collaborative-filtering techniques. A set of an item's nearest neighbor items based on the set of item-item affinities is determined. A set of user feature-item affinities for the plurality of items and a set of user features is determined based on least squared regression. A set of a user feature's nearest neighbor items is determined based in part on the set of user feature-item affinities. Compatible affinity weights for nearest neighbor items of each item and each user feature are determined and stored. Based on user features of a particular user and items a particular user has consumed, a set of nearest neighbor items comprising nearest neighbor items for user features of the user and items the user has consumed are identified as a set of candidate items, and affinity scores of candidate items are determined.

Abstract: Content items are selected to be displayed on a portal page in such a way as to maximize a performance metric such as click-through rate. Problems relating to content selection are addressed, such as changing content pool, variable performance metric, and delay in receiving feedback on an item once the item has been displayed to a user. An adaptation of priority-based schemes for the multi-armed bandit problem are used to project future trends of data. The adaptation introduces experiments concerning a future time period into the calculation, which increases the set of data on which to solve the multi-armed bandit problem. Also, a Bayesian explore/exploit method is formulated as an optimization problem that addresses all of the issues of content item selection for a portal page. This optimization problem is modified by Lagrange relaxation and normal approximation, which allow computation of the optimization problem in real time.

Abstract: An improved system and method for pricing of overlapping impression pools of online advertisement impressions for advertising demand is provided. An inventory of online advertisement impressions may be grouped in impression pools according to attributes of the advertisement impressions and advertisers' requests for impressions targeting specific attributes may be received. An optimal price may be computed for each of the impression pools of the inventory of online advertisement impressions using dual values of an optimization program. The values of a dual variable for prices of impression pools on the supply constraints of an objective function for allocating the impression pools may be extracted and iteratively increased on those impression pools which have a dual value greater than the book rate value.

Abstract: Techniques are described herein for determining whether to provide an advertisement to a user of a social network. The determination is based on a click probability and a social network value for the user. The click probability indicates a likelihood of the user to select the advertisement if provided to the user via the social network. The social network value is based on a subscription probability of the user and further based on subscription probabilities of other users in the social network that are included in an affinity set of the user. Each subscription probability indicates a likelihood of a respective user to subscribe to a paid service with respect to the social network.

Abstract: According to some example embodiments, a method includes calculating learning values associated with a plurality of listings, at least one of said learning values associated with one of said listings representing a value based, at least in part, on a probability distribution of selections of said listing. The method further includes applying said learning values to ranking scores associated with said listings to provide an updated ranking, and electronically auctioning advertising inventory to purchasers associated with said listings based, at least in part, on said updated ranking.

Abstract: Methods and systems are provided for click through rate prediction and advertisement selection in online advertising. Methods are provided in which output information from a feature-based machine learning model is utilized. The output information includes predicted click through rate information. The output information is used to form a matrix. The matrix is modeled using a latent variable model. Machine learning techniques can be used in determining values for unfilled cells of one or more model matrices. The latent variable model can be used in determining predicted click through rate information, and in advertisement selection in connection with serving opportunities.