Method and system for adaptive online updating of ad related models

Abstract

The present teaching relates to generating an updated model related to advertisement selection. In one example, a request is obtained for updating a model to be utilized for selecting an advertisement. A plurality of copies of the model is generated. The model is pre-selected based on a performance metric related to advertisement selection. Based on each of the plurality of copies, a candidate model is created by modifying one or more parameters of the copy of the model to create a plurality of candidate models. One of the plurality of candidate models is selected based on the performance metric. The steps of generating, creating, and selecting are repeated until a predetermined condition is met. The model is updated with the latest selected candidate model when the predetermined condition is met.

We claim: 
     
         1 . A method for selecting an advertisement, the method comprising:
 receiving, from a user, an advertisement request;   collecting bids from advertisers;   obtaining, based on the bids, one or more advertisements associated with one or more of the advertisers;   selecting, based on a model, at least one of the one or more of the advertisements, wherein the model is trained by:
 creating a plurality of candidate models, 
 selecting one of the plurality of candidate models based on a performance metric associated with each of the plurality of candidate models, and 
 designating the selected candidate model as the model when a condition is met; and 
   providing the selected at least one advertisement to the user.   
     
     
         2 . The method of  claim 1 , wherein the selecting the at least one of the one or more of the advertisements is further based on a user profile of the user which is retrieved based on the advertisement request. 
     
     
         3 . The method of  claim 1 , wherein the creating the plurality of candidate models is by modifying one or more parameters of each of a plurality of copies of the model. 
     
     
         4 . The method of  claim 3 , wherein the modifying one or more parameters of each of the plurality of copies of the model is based on one or more scale factors which are based on one or more temporal changes. 
     
     
         5 . The method of  claim 4 , wherein the one or more temporal changes indicate a shift from a first time period associated with a first amount of new advertisements to a second time period associated with a second amount, different from the first amount, of new advertisements. 
     
     
         6 . The method of  claim 4 , wherein the one or more temporal changes indicate a change of a weight associated with a type of learning of new advertisements associated with the training of the model and are detected from streaming ad-related data. 
     
     
         7 . The method of  claim 1 , wherein the condition is based on a level of convergence of the training of the model. 
     
     
         8 . A non-transitory, computer-readable medium having information recorded thereon for selecting an advertisement, wherein the information, when read by a machine, causes the machine to perform operations comprising:
 receiving, from a user, an advertisement request;   collecting bids from advertisers;   obtaining, based on the bids, one or more advertisements associated with one or more of the advertisers;   selecting, based on a model, at least one of the one or more of the advertisements, wherein the model is trained by:
 creating a plurality of candidate models, 
 selecting one of the plurality of candidate models based on a performance metric associated with each of the plurality of candidate models, and 
 designating the selected candidate model as the model when a condition is met; and 
   providing the selected at least one advertisement to the user.   
     
     
         9 . The medium of  claim 8 , wherein the selecting the at least one of the one or more of the advertisements is further based on a user profile of the user which is retrieved based on the advertisement request. 
     
     
         10 . The medium of  claim 8 , wherein the creating the plurality of candidate models is by modifying one or more parameters of each of a plurality of copies of the model. 
     
     
         11 . The medium of  claim 10 , wherein the modifying one or more parameters of each of the plurality of copies of the model is based on one or more scale factors which are based on one or more temporal changes. 
     
     
         12 . The medium of  claim 11 , wherein the one or more temporal changes indicate a shift from a first time period associated with a first amount of new advertisements to a second time period associated with a second amount, different from the first amount, of new advertisements. 
     
     
         13 . The medium of  claim 11 , wherein the one or more temporal changes indicate a change of a weight associated with a type of learning of new advertisements associated with the training of the model and are detected from streaming ad-related data. 
     
     
         14 . The medium of  claim 8 , wherein the condition is based on a level of convergence of the training of the model. 
     
     
         15 . A system for selecting an advertisement, the system comprising:
 memory storing computer program instructions; and   one or more processors that, in response to executing the computer program instructions, effectuate operations comprising:   receiving, from a user, an advertisement request;   collecting bids from advertisers;   obtaining, based on the bids, one or more advertisements associated with one or more of the advertisers;   selecting, based on a model, at least one of the one or more of the advertisements, wherein the model is trained by:
 creating a plurality of candidate models, 
 selecting one of the plurality of candidate models based on a performance metric associated with each of the plurality of candidate models, and 
 designating the selected candidate model as the model when a condition is met; and 
   providing the selected at least one advertisement to the user.   
     
     
         16 . The system of  claim 15 , wherein the selecting the at least one of the one or more of the advertisements is further based on a user profile of the user which is retrieved based on the advertisement request. 
     
     
         17 . The system of  claim 15 , wherein the creating the plurality of candidate models is by modifying one or more parameters of each of a plurality of copies of the model. 
     
     
         18 . The system of  claim 17 , wherein the modifying one or more parameters of each of the plurality of copies of the model is based on one or more scale factors which are based on one or more temporal changes. 
     
     
         19 . The system of  claim 18 , wherein the one or more temporal changes indicate a shift from a first time period associated with a first amount of new advertisements to a second time period associated with a second amount, different from the first amount, of new advertisements. 
     
     
         20 . The system of  claim 18 , wherein the one or more temporal changes indicate a change of a weight associated with a type of learning of new advertisements associated with the training of the model and are detected from streaming ad-related data.