Amazon Technologies, Inc.

Jim Huang

Yu Gan

Fabian Lutz-Frank Wauthier

Jonathan Lee Burstein

Noor-E-Gagan Singh

Rodolphe Jenatton

Paula Marie Despins

Weam Abu Zaki

Xuan Wu

Kelly Paulson

Patrick Jing Ye

Kyle Tiffany

Cody Barnhart

Cédric Philippe Archambeau

Teresa Chen

Strategies for an objective associated with an offering set are obtained. A strategy assigns respective selection probabilities of receiving content associated with the offering set to users of a user population. Strategy optimization iterations are performed with respect to a sub-sample of the population and a subset of the strategies. In a given iteration, weights assigned to the strategies are used to determine aggregated selection probabilities for users, content pertaining to the offering set is presented to users selected based on the aggregated probabilities, and the weights are adjusted based on feedback metrics and an exploration-exploitation tradeoff parameter. Based on weights updated in the iterations, content associated with the offering set is presented to users which were not in the sub-sample.

Optimization techniques for content presentation strategies

Systems, methods, and computer-readable media are disclosed for automated server-based content delivery. In one embodiment, an example method may include determining campaign information for a content delivery campaign, the campaign information comprising a first allocation value for first content, and a second allocation value for second content, determining first observed data comprising a first user response rate for the first content and a second user response rate for the second content, determining an exponentiated gradient algorithm for the content delivery campaign, and determining a reallocation amount to reallocate a portion of the first allocation value to the second allocation value using the exponentiated gradient algorithm based at least in part on the first observed data, wherein the reallocation amount maximizes an output of the exponentiated gradient algorithm.

Automated server-based content delivery

Systems, methods, and computer-readable media are disclosed for detecting user interactions and delivering content using interaction metrics. In one embodiment, an example method may include receiving a bid request for an available content delivery slot, the bid request comprising context information, determining first candidate content for the available content delivery slot, and determining a first base bid value for the first candidate content. Example methods may include determining a predicted conversion rate for an impression of the first candidate content served at the available content delivery slot, determining an estimated revenue for serving the impression at the available content delivery slot, determining a first bid modifier using the predicted conversion rate and the estimated revenue, and sending a response to the bid request comprising a first bid amount, wherein the first bid amount is based at least in part on the first base bid value and the first bid modifier.

Detecting user interaction and delivering content using interaction metrics

Systems, methods, and computer-readable media are disclosed for detecting viewability and delivering content using viewability metrics. In one embodiment, an example method may include receiving a bid request for an available content delivery slot, where the bid request comprises context information, determining first candidate content for the available content delivery slot, and determining a first base bid value for the first candidate content. Example methods may include determining an estimated view rate using the context information, determining a first bid modifier using the estimated view rate; and sending a response to the bid request comprising a first bid amount. The first bid amount may be based at least in part on the first base bid value and the first bid modifier.

Detecting viewability and delivering content using viewability metrics

Systems, methods, and computer-readable media are disclosed for maximizing quantifiable user interaction via modification of adjustable parameters. In one embodiment, an example method may include determining a first output to maximize, where the first output is a function of a first adjustable parameter and a second adjustable parameter, determining first data comprising a first actual value of the first output when the first adjustable parameter is set to a first value and the second adjustable parameter is set to a second value, and determining a first predictive model that generates a first predicted value of the first output. Example methods may include determining, using the first predictive model, a third value for the first adjustable parameter and a fourth value for the second adjustable parameter to maximize the first predicted value, and sending the third value and the fourth value.

Maximizing quantifiable user interaction via modification of adjustable parameters

Systems, methods, and computer-readable media are disclosed for optimization of real-time probabilistic model evaluation for online advertising. In one embodiment, a system may receive a bid request for an ad slot, and may receive cached user information with a first ad and a first ad score, and a second ad and a second ad score. The system may generate a first estimated probability of conversion associated with presentation of the first ad to the user, and may generate a second estimated probability of conversion associated with presentation of the second ad to the user based at least in part on the context information, the second ad identifier, and the second ad score. The system may select either the first ad or the second ad for which to generate a bid amount based at least in part on the first estimated probability of conversion and the second estimated probability of conversion.

Growing community of inventors

Seattle, WA, United States of America

Jim Huang