Online serving threshold and delivery policy adjustment

Abstract

The present invention provides techniques for use in association with online advertising, relating to use of serving thresholds, associated with predicted click through rates, and delivery policies, associated with advertising inventory serving and distribution. An offline-trained machine learning-based model may be utilized in advertising serving decision-making in connection with serving opportunities, However, serving thresholds and delivery policies, for use in association with the model in serving decision-making, may he adjusted online, such as in real-time or near real-time, based on information obtained online affecting factors such as predicted click through rates and advertising inventory distribution.

1 . A method comprising:
 during an offline period, determining a set of serving thresholds for online advertisement serving;   during an offline period, determining a set of delivery policies for distribution of advertising inventory across advertisement serving opportunities;   during an online period, adjusting at least one of the set of serving thresholds to determine at least one adjusted serving threshold, and adjusting at least one of the set of delivery policies to determine at least one adjusted delivery policy; and   during an online period, modifying serving of online advertisements in response to advertisement serving opportunities based on a machine-learning model that is trained on advertising information associated with serving advertisements during an offline period and updated during an online period, and further based at least in part on the at least one adjusted serving threshold and the at least one adjusted delivery policy, adjusted during the online period.   
     
     
         2 . The method of  claim 1 , wherein the advertising information includes user behavior information. 
     
     
         3 . The method of  claim 1 , wherein the advertising information includes advertisement performance information. 
     
     
         4 . The method of  claim 1 , wherein the advertising information includes advertisement inventory information. 
     
     
         5 . The method of  claim 1 , wherein the advertising information includes advertisement distribution information. 
     
     
         6 . The method of  claim 1 , wherein the advertising information is associated with a node of a hierarchical behavioral targeting taxonomy. 
     
     
         7 . The method of  claim 1 , further comprising periodically training the machine-learning model with the advertising information. 
     
     
         8 . The method of  claim 1 , further comprising adjusting at least one of the set of serving thresholds and at least one of the set of delivery policies based on a news event occurring during an online period. 
     
     
         9 . The method of  claim 1 , further comprising adjusting at least one of the set of serving thresholds and at least one of the set of delivery policies based on a chronological event occurring during an online period. 
     
     
         10 . The method of  claim 1 , further comprising:
 during an offline period, training the machine-learning model with the advertising information; and   during an online period, adjusting at least one of the set of serving thresholds and at least one of the set of delivery policies based on a news event or a chronological event occurring during an online period and prior to re-training the machine-learning model during a next offline period based on advertising information generated during the online period and associated with the new event or with the chronological event.   
     
     
         11 . A system comprising one or more computers and one or more storage devices on which are stored instructions that are operable, when executed by the one or more computers, to cause the one or more computers to perform operations comprising:
 during an offline period, determining a set of serving thresholds for online advertisement serving;   during an offline period, determining a set of delivery policies for distribution of advertising inventory across advertisement serving opportunities;   during an online period, adjusting at least one of the set of serving thresholds to determine at least one adjusted serving threshold, and adjusting at least one of the set of delivery policies to determine at least one adjusted delivery policy; and   during an online period, modifying, based on a machine-learning model that is trained on advertising information associated with serving advertisements during an offline period and updated during an online period, serving of online advertisements in response to serving opportunities based at least in part on the at least one adjusted serving threshold and the at least one adjusted delivery policy, adjusted during the online period.   
     
     
         12 . The system of  claim 11 , the operations further comprising:
 during an offline period, training the machine-learning model with the advertising information; and   during an online period, adjusting at least one of the set of serving thresholds and at least one of the set of delivery policies based on a news event or a chronological event occurring during an online period and prior to re-training the machine-learning model during a next offline period based on advertising information generated during the online period and associated with the new event or with the chronological event.   
     
     
         13 . The system of  claim 11 , wherein the advertising information includes user behavior information. 
     
     
         14 . The system of  claim 11 , wherein the advertising information includes advertisement performance information. 
     
     
         15 . The system of  claim 11 , wherein the advertising information includes advertisement inventory information. 
     
     
         16 . The system of  claim 11 , wherein the advertising information includes advertisement distribution information. 
     
     
         17 . The system of  claim 11 , wherein the advertising information is associated with a node of a hierarchical behavioral targeting taxonomy. 
     
     
         18 . The system of  claim 11 , the operations further comprising adjusting at least one of the set of serving thresholds and at least one of the set of delivery policies based on a news event or a chronological event occurring during an online period. 
     
     
         19 . A non-transitory computer readable medium or media containing instructions for executing a method for use in association with an online advertising marketplace, the method comprising:
 during an offline period, determining a set of serving thresholds for online advertisement serving;   during an offline period, determining a set of delivery policies for distribution of advertising inventory across advertisement serving opportunities;   during an online period, adjusting at least one of the set of serving thresholds to determine at least one adjusted serving threshold, and adjusting at least one of the set of delivery policies to determine at least one adjusted delivery policy; and   during an online period, modifying serving of online advertisements in response to advertisement serving opportunities based on a machine-learning model that is trained on advertising information associated with serving advertisements during an offline period and updated during an online period, and further based at least in part on the at least one adjusted serving threshold and the at least one adjusted delivery policy, adjusted during the online period.   
     
     
         20 . The non-transitory computer readable medium of  claim 19 , the method further comprising:
 during an offline period, training the machine-learning model with the advertising information; and   during an online period, adjusting at least one of the set of serving thresholds and at least one of the set of delivery policies based on a news event or a chronological event occurring during an online period and prior to re-training the machine-learning model during a next offline period based on advertising information generated during the online period and associated with the new event or with the chronological event.   
     
     
         21 . The non-transitory computer readable medium of  claim 19 , the operations further comprising adjusting at least one of the set of serving thresholds and at least one of the set of delivery policies based on a news event or a chronological event occurring during an online period.