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: A method is disclosed for limiting latency in filling a display opportunity in an ad exchange including: constructing an exchange graph comprising nodes representing a plurality of publishers and advertisers, the exchange graph also including a plurality of directed edges that represent bilateral business agreements connecting the nodes; receiving an opportunity for displaying an ad to a user, wherein the opportunity is associated with a publisher node; receiving ads from the advertisers from which to choose to fill the opportunity; determining whether a threshold total number of ads (T) is surpassed by the received ads; and randomly downsampling the number of ads from each of at least some of the advertisers when the threshold total number of ads (T) is surpassed by the received ads to reduce the total number of ads to a target number of ads (S) that reduces overall latency in determining which of sampled ads will fill the opportunity.

Abstract: A method is disclosed for optimizing ad selection in an exchange having intermediate ad-networks including: constructing an exchange graph having nodes representing publishers, advertisers, and intermediate ad-network entities, and including directed edges that represent bilateral business agreements connecting the nodes; receiving an opportunity for displaying an ad to a user that is associated with a publisher node and includes properties that are targetable by supply predicates, wherein a supply predicate is a function whose inputs include properties of the user; receiving ads that are available for display to the user associated with respective advertiser nodes and that include properties that are targetable by demand predicates, wherein a demand predicate is a function whose inputs include properties of one or more of the plurality of ads; computing a thinned graph by enforcing the supply predicates in the nodes and edges of the graph; and producing a list of ads and corresponding paths that exist throug

Abstract: Disclosed is a system to price usage of a user-action Probability estimation system provided by an advertising exchange system. A bid from each bidder in an auction for an advertising opportunity is presented in a computer. The bidders comprise a first group of bidders that utilize the Probability estimation system and a second group of bidders that do not utilize the Probability estimation system. The bids are processed by determining a first equilibrium bid for a first bidder as a member of the first group. The bids are further processed by determining a second equilibrium bid for the first bidder as a member of the second group. The system then utilizes the first equilibrium bid and the second equilibrium bid to determine a value of utilizing the Probability estimation system.

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: Methods and systems are provided for use in connection with an online advertising exchange. Methods are provided in which an auction format is utilized. However, generally, selection and pricing is rate-based and based on near-term forecasting. As such, the selection and pricing does not match specific, immediately available serving opportunities and advertisement inventory. Methods are provided which are in some ways analogous to providing futures—like mini-contracts on an advertising exchange, in which selection and pricing are decoupled from fulfillment of the mini-contracts in the form of serving of advertisements. This can allow auction format advantages while yet allowing forecasting, offline processing and relatively inexpensive operation costs.

Abstract: A system and method to distribute computation for an exchange in which advertisers buy online advertising space from publishers. The exchange maintains submarkets, each containing a subset of the ad calls supplied by publishers and a subset of the offers and budgets representing demand from advertisers. Portfolio optimization techniques allocate the supply of ad calls from publishers over the submarkets, with the goal of maximizing profits for publishers while limiting the volatility of those profits. Portfolio optimization techniques allocate the demand from advertisers over the submarkets, with the goal of maximizing return on investment for advertisers. The exchange re-allocates supply and demand over submarkets periodically. Also, periodically, the most effective submarkets are replicated and the least effective submarkets are eliminated.