Detecting and correcting potential errors in user behavior

Abstract

A computing system is described that predicts a future action to be taken by a user of a computing device and determines, based on contextual information associated with the computing device, a current action being taken by the user. The computing system determines, based on the current action, a degree of likelihood of whether the user will be able to take the future action and predicts, based on the degree of likelihood, that the user will not be able to take the future action. The computing system sends information to the computing device indicating that the current action being taken by the user will lead to the user not being able to take the future action.

What is claimed is: 
     
         1 . A method implemented by one or more processors, the method comprising:
 determining, based on user information associated with a user of a computing device, a future action to be taken by the user of the computing device;   during a first range of time:
 determining, based on a first instance of contextual information associated with the computing device, a first current action being taken by the user at a first time within the first range of time; 
 determining, based on the first current action being taken by the user at the first time, a first degree of likelihood of whether the user will be able to take the future action; and 
 refraining from sending, to the computing device of the user and based on the first degree of likelihood, any information proactively notifying the user that the first current action being taken by the user will lead to the user not being able to take the future action; and 
   during a second range of time that is subsequent to the first range of time and that is prior to the future action being taken by the user:
 determining, based on a second instance the contextual information associated with the computing device, a second current action being taken by the user at the second time within the second range of time; 
 determining, based on the second current action being taken by the user at the second time, a second degree of likelihood of whether the user will be able to take the future action; and 
 sending, to the computing device of the user, and based on the second degree of likelihood, information that proactively notifies the user that the second current action being taken by the user will lead to the user not being able to take the future action and that proactively notifies the user of one or more corrective actions,
 wherein sending the information causes the computing device to automatically present the information to the user via the computing device of the user and during the second range of time. 
 
   
     
     
         2 . The method of  claim 1 , wherein the first instance of the contextual information includes search history information of the user of the computing device during the first range of time, and wherein the first current action includes searching for navigational directions to a second location that is different from a first location associated with the future action. 
     
     
         3 . The method of  claim 2 , wherein the first instance of the contextual information does not include movement information of the computing device of the user during the first range of time. 
     
     
         4 . The method of  claim 3 , wherein the second instance of the contextual information includes additional movement information of the computing device of the user, and wherein the second current action being taken by the user during the second range of time includes travelling to the second location that is different from the first location associated with the future action. 
     
     
         5 . The method of  claim 4 ,
 wherein determining the first degree of likelihood of whether the user will be able to take the future action comprises:
 applying, as input across a machine learning model to generate first output, the search history information of the user of the computing device during the first range of time; and 
 determining the first degree of likelihood of whether the user will be able to take the future action based on the first output; and 
   wherein determining the second degree of likelihood of whether the user will be able to take the future action comprises:
 applying, as input across the machine learning model to generate second output, the additional movement information of the computing device of the user during the second range of time; and 
 determining the second degree of likelihood of whether the user will be able to take the future action based on the second output. 
   
     
     
         6 . The method of  claim 2 , wherein the second location and the first location associated with future action are one of: different respective locations of a given transportation service, different respective offices of a single commercial establishment, or different respective cities of different states. 
     
     
         7 . The method of  claim 2 , wherein the second instance of the contextual information includes additional search history information of the user of the computing device during the second range of time, and wherein the second current action includes searching for the navigational directions to the wrong location that is different from the location associated with the future action. 
     
     
         8 . The method of  claim 7 ,
 wherein determining the first degree of likelihood of whether the user will be able to take the future action comprises:
 applying, as input across a machine learning model to generate first output, the search history information of the user of the computing device during the first range of time; and 
 determining the first degree of likelihood of whether the user will be able to take the future action based on the first output; and 
   wherein determining the second degree of likelihood of whether the user will be able to take the future action comprises:
 applying, as input across the machine learning model to generate second output, the additional search history information of the user of the computing device during the second range of time; and 
 determining the second degree of likelihood of whether the user will be able to take the future action based on the second output. 
   
     
     
         9 . The method of  claim 1 ,
 wherein the first instance of the contextual information includes movement information of the computing device of the user during the first range of time,   wherein the first current action includes travelling to a second location that is different from a first location associated with the future action,   wherein the second instance of the contextual information includes additional movement information of the computing device of the user during the second range of time, and   wherein the second current action also includes travelling to the second location that is different from the first location associated with the future action.   
     
     
         10 . The method of  claim 9 ,
 wherein determining the first degree of likelihood of whether the user will be able to take the future action comprises:
 applying, as input across a machine learning model to generate first output, the movement information of the computing device of the user during the first range of time; and 
 determining the first degree of likelihood of whether the user will be able to take the future action based on the first output; and 
   wherein determining the second degree of likelihood of whether the user will be able to take the future action comprises:
 applying, as input across the machine learning model to generate second output, the additional movement information of the computing device of the user during the second range of time; and 
 determining the second degree of likelihood of whether the user will be able to take the future action based on the second output. 
   
     
     
         11 . The method of  claim 1 , the method further comprising:
 in response to determining the second degree of likelihood satisfies a threshold:
 determining one or more of the corrective actions to include in the information sent to the computing device of the user. 
   
     
     
         12 . The method of  claim 11 , wherein the future action to be taken by the user of the computing device is associated with a first respective office of a single commercial establishment, and wherein determining one or more of the corrective actions comprises:
 submitting, to a search system, a query for a plurality of respective offices of the single commercial establishment; and   selecting a given one of the plurality of the respective office based on a current location of the computing device of the user.   
     
     
         13 . A computing system, comprising:
 at least one processor; and   at least one module operable by the at least one processor to:
 determine, based on user information associated with a user of a computing device, a future action to be taken by the user of the computing device; 
 during a first range of time:
 determine, based on a first instance of contextual information associated with the computing device, a first current action being taken by the user at a first time within the first range of time; 
 determine, based on the first current action being taken by the user at the first time, a first degree of likelihood of whether the user will be able to take the future action; and 
 refrain from sending, to the computing device of the user, and based on the first degree of likelihood, any information proactively notifying the user that the first current action being taken by the user will lead to the user not being able to take the future action; and 
 
 during a second range of time that is subsequent to the first range of time and that is prior to the future action being taken by the user:
 determine, based on a second instance the contextual information associated with the computing device, a second current action being taken by the user at the second time within the second range of time; 
 determine, based on the second current action being taken by the user at the second time, a second degree of likelihood of whether the user will be able to take the future action; and 
 send, to the computing device of the user, and based on the second degree of likelihood, information that proactively notifies the user that the second current action being taken by the user will lead to the user not being able to take the future action and that proactively notifies the user of one or more corrective actions, 
 wherein sending the information causes the computing device to automatically present the information to the user via the computing device of the user and during the second range of time. 
 
   
     
     
         14 . The computing system of  claim 13 ,
 wherein the first instance of the contextual information includes search history information of the user of the computing device during the first range of time,   wherein the first instance of the contextual information does not include movement information of the computing device of the user during the first range of time.   wherein the first current action includes searching for navigational directions to a wrong location that is different from a location associated with the future action,   wherein the second instance of the contextual information includes additional movement information of the computing device of the user during the second range of time, and   wherein the second current action includes travelling to a second location that is different from a first location associated with the future action.   
     
     
         15 . The computing system of  claim 14 ,
 wherein in determining the first degree of likelihood of whether the user will be able to take the future action the at least one module is further operable by the at least one processor to:
 apply, as input across a machine learning model to generate first output, the search history information of the user of the computing device during the first range of time; and 
 determine the first degree of likelihood of whether the user will be able to take the future action based on the first output; and 
   wherein in determining the second degree of likelihood of whether the user will be able to take the future action the at least one module is further operable by the at least one processor to:
 apply, as input across the machine learning model to generate second output, the additional movement information of the computing device of the user during the second range of time; and 
 determine the second degree of likelihood of whether the user will be able to take the future action based on the second output. 
   
     
     
         16 . The computing system of  claim 13 ,
 wherein the first instance of the contextual information includes movement information of the computing device of the user during the first range of time,   wherein the first current action includes travelling to a second location that is different from a first location associated with the future action,   wherein the second instance of the contextual information includes additional movement information of the computing device of the user during the second range of time, and   wherein the second current action also includes travelling to the second location that is different from the first location associated with the future action.   
     
     
         17 . The computing system of  claim 16 ,
 wherein in determining the first degree of likelihood of whether the user will be able to take the future action the at least one module is further operable by the at least one processor to:
 apply, as input across a machine learning model to generate first output, the movement information of the computing device of the user during the first range of time; and 
 determine the first degree of likelihood of whether the user will be able to take the future action based on the first output; and 
   wherein in determining the second degree of likelihood of whether the user will be able to take the future action the at least one module is further operable by the at least one processor to:
 apply, as input across the machine learning model to generate second output, the additional movement information of the computing device of the user during the second range of time; and 
 determine the second degree of likelihood of whether the user will be able to take the future action based on the second output. 
   
     
     
         18 . A non-transitory computer-readable storage medium comprising instructions that, when executed, cause at least one processor of a computing system to:
 determine, based on user information associated with a user of a computing device, a future action to be taken by the user of the computing device;   during a first range of time:
 determine, based on a first instance of contextual information associated with the computing device, a first current action being taken by the user at a first time within the first range of time; 
 determine, based on the first current action being taken by the user at the first time, a first degree of likelihood of whether the user will be able to take the future action; and 
 refrain from sending, to the computing device of the user, and based on the first degree of likelihood, any information proactively notifying the user that the first current action being taken by the user will lead to the user not being able to take the future action; and 
   during a second range of time that is subsequent to the first range of time and that is prior to the future action being taken by the user:
 determine, based on a second instance the contextual information associated with the computing device, a second current action being taken by the user at the second time within the second range of time; 
 determine, based on the second current action being taken by the user at the second time, a second degree of likelihood of whether the user will be able to take the future action; and 
 send, to the computing device of the user, and based on the second degree of likelihood, information that proactively notifies the user that the second current action being taken by the user will lead to the user not being able to take the future action and that proactively notifies the user of one or more corrective actions, 
   
       wherein sending the information causes the computing device to automatically present the information to the user via the computing device of the user during the second range of time.