Method and system for semi-supervised semantic task management from semi-structured heterogeneous data streams
Embodiments of the present invention are directed to a computer-implemented machine-learning method and system for automatically creating and updating tasks by reading signals from external data sources and understanding what users are doing. Embodiments of the present invention are directed to a computer-implemented machine-learning method and system for automatically completing tasks by reading signals from external sources and understanding when an existing task has been executed. Tasks created are representable and explainable in a human readable format that can be shown to users and used to automatically fill productivity applications including but not limited to task managers, to-do lists, project management, time trackers, and daily planners. Tasks created are representable in a way that can be interpreted by a machine such as a computer system or an artificial intelligence so that external systems can be delegated or connected to the system.
1. A computer-implemented machine-learning method comprising:
detecting a trigger event from a user-selected source, in which the trigger event comprises a heterogeneous message representation used to import a signal from a third-party system so that the signal can be processed for a purpose of inferring intentions;
analyzing the trigger event to determine whether to associate the trigger event with one of several ongoing activities or with a newly created activity, each activity consisting of a goal achievable with a sequence of steps between users;
processing the trigger event using machine-learning to determine an intention by combining metadata associated with the trigger event with structured text of the trigger event and natural text of the trigger event, the intention directed to completion of a goal in the activity;
creating a task associated with the activity associated with the triggering event if the intention indicates a task that a user can complete; and
creating an expectation associated with the activity associated with the triggering event if the intention indicates a task that the user must wait for someone other than the user to complete,
wherein analyzing the trigger event to determine whether to associate the trigger event with one of several existing activities or with a newly created activity comprises clustering trigger events already associated with ongoing activities and computing the geometrical distance between the trigger event being processed and the resulting clusters.
2. The computer-implemented method of claim 1 , further comprising:
detecting a second trigger event;
processing the second trigger event using machine-learning to determine that an action has been performed, in which an action is a suggested way of executing and completing the task;
marking the task complete if the action indicates that the task has been completed by the user; and
marking the expectation as complete if the action indicates that the task has been completed by someone other than the user.
3. The computer-implemented method of claim 1 , further comprising:
in response to determining the trigger event is associated with an existing activity, updating the existing activity with the trigger event;
in response to determining the trigger event is not associated with an existing activity, using a machine-learning classifier to determine whether to create a new activity with the trigger event or disregard the trigger event as being irrelevant;
creating a new activity if the machine-learning classifier determines a new activity should be created in response to the trigger event; and
disregarding the trigger event as being irrelevant if the machine-learning classifier determines a new activity should not be created in response to the trigger event.
4. The method of claim 3 , further comprising using machine-learning to determine a plurality of tasks required to complete the activity.
5. The computer-implemented method of claim 1 , in which the created task is representable and explainable in a human readable format displayable to the user and used to automatically fill a productivity application.
6. The computer-implemented method of claim 5 , in which the productivity application is selected from a group consisting of: task managers, to-do lists, project management, time trackers, and daily planners.
7. The computer-implemented method of claim 1 , in which the created task created is representable in a way that can be interpreted by a delegated system comprising an external computer system or an external artificial intelligence.
8. The computer-implemented method of claim 7 , in which the delegated system uses the task to automate one or more features of the delegated system.
9. The computer-implemented method of claim 7 , in which the delegated system suggests an action to complete the created task or expectation.
10. The method of claim 1 further comprising creating a second task by predicting a subsequent event associated with the activity associated with the triggering event.
11. The method of claim 10 in which predicting a subsequent event associated with the activity associated with the triggering comprises predicting a subsequent event using a recurrent network.
12. A computer system comprising:
a processor;
a network adapter coupled to the processor for receiving signals from a third-party system and sending signals to the third-party system;
a computer-readable memory coupled to the processor, the computer-readable memory programmed with computer-executable instructions that, when executed by the processor, cause the computer system to perform the steps of:
detecting a trigger event from a user-selected source, in which the trigger event comprises a heterogeneous message representation used to import a signal from the third-party system so that the signal can be processed for a purpose of inferring intentions;
analyzing the trigger event to determine whether to associate the trigger event with one of several ongoing activities or with a newly created activity, an activity consisting of a goal achievable with a sequence of steps between users;
processing the trigger event using machine-learning to determine an intention by combining metadata associated with the trigger event with structured text of the trigger event and natural text of the trigger event, the intention directed to completion of a goal in the activity to which the user actively contributes;
creating a task associated with the activity associated with the triggering event if the intention indicates a task that a user can complete; and
creating an expectation associated with the activity associated with the triggering event if the intention indicates a task that the user must wait for someone other than the user to complete
in which the step of analyzing the trigger event to determine whether to associate the trigger event with one of several existing activities or with a newly created activity comprises clustering trigger events already associated with ongoing activities and computing the geometrical distance between the trigger event being processed and the resulting clusters.
13. The computer system of claim 12 , further programmed to perform the steps of:
detecting a second trigger event;
processing the second trigger event using machine-learning to determine that an action has been performed, in which an action is a suggested way of executing and completing the task;
marking the task complete if the action indicates that the task has been completed by the user; and
marking the expectation as complete if the action indicates that the task has been completed by someone other than the user.
14. The computer system of claim 12 , further programmed to perform the steps of:
in response to determining the trigger event is associated with an existing activity, updating the existing activity with the trigger event;
in response to determining the trigger event is not associated with an existing activity, using a machine-learning classifier to determine whether to create a new activity with the trigger event or disregard the trigger event as being irrelevant;
creating a new activity if the machine-learning classifier determines a new activity should be created in response to the trigger event; and
disregarding the trigger event as being irrelevant if the machine-learning classifier determines a new activity should not be created in response to the trigger event.
15. The computer system of claim 14 , further programmed to use machine-learning to determine a plurality of tasks required to complete the activity.
16. The computer system of claim 12 , in which the created task is representable and explainable in a human readable format that displayable to the user and used to automatically fill a productivity application.
17. The computer system of claim 16 , in which the productivity application is selected from a group consisting of: task managers, to-do lists, project management, time trackers, and daily planners.
18. The computer system of claim 12 , in which the created task created is representable in a way interpretable by a delegated system comprising an external computer system or an external artificial intelligence.
19. The computer system of claim 18 , in which the delegated system uses the task to automate one or more features of the delegated system.
20. The computer system of claim 18 , in which the delegated system suggests an action to complete the created task or expectation.