Cognitive robotics analyzer

• US 9,805,306 B1
• Filed: 11/23/2016
• Issued: 10/31/2017
• Est. Priority Date: 11/23/2016
• Status: Active Grant

• Current Assignee: Accenture Global Solutions Limited (Accenture PLC)
• Original Assignee: Accenture Global Solutions Limited (Accenture PLC)
• Inventors: Bataller, Cyrille, Schiopu, Vitalie, Jacquot, Adrien, Torres, Sergio Ral Duarte, Hall, Simon
• Primary Examiner(s): Misir, Dave
• Application Number:
  US15/360,535
• Time in Patent Office: 342 Days
• Field of Search: 706 21
• CPC Class Codes: G06F 3/0487 using specific features provided by the input device, e.g. functions controlled by the rotation of a mouse with dual sensing arrangements, or of the nature of the input device, e.g. tap gestures based on pressure sensed by a digitiser
• CPC Class Codes: G06N 20/00 Machine learning
• CPC Class Codes: G06N 3/08 Learning methods
• CPC Class Codes: G06N 5/022 Knowledge engineering; Knowledge acquisition
• CPC Class Codes: G06Q 10/10 Office automation; Time management
• CPC Class Codes: G06Q 30/0201 Market modelling; Market analysis; Collecting market data
• CPC Class Codes: G06Q 30/0202 Market predictions or forecasting for commercial activities
• CPC Class Codes: H04L 43/14 using software, i.e. software packages network security related monitoring H04L63/1408
• CPC Class Codes: H04L 67/306 User profiles
• CPC Class Codes: H04L 67/535 Tracking the activity of the user network monitoring arrangements H04L43/00; recording of computer activity G06F11/34

Abstract

Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for a cognitive robotics analyzer are disclosed. In one aspect, a method includes the actions of receiving, for each user characteristic of a plurality of user characteristics, first data that identifies one or more first actions that perform a first process and second data that identifies one or more second actions that perform a second process that is labeled as similar to the first process. The actions further include training a predictive model. The actions further include receiving data that identifies actions performed by a user. The actions further include applying the predictive model to one or more of the actions. The actions further include classifying a process performed by the one or more actions as similar to a particular process. The actions further include associating the user with the particular user characteristic.

Claims

1. A computer-implemented method comprising:

receiving first action data that identifies one or more first actions that were initiated by a first user and second action data that identifies one or more second actions that were initiated by a second user;

receiving first process data that identifies a first process performed by the one or more first actions and second process data that identifies a second process performed by the one or more second actions;

receiving process similarity data that indicates that the first process is similar to the second process;

receiving user characteristic data that identifies a common demographic, geographic, or employment attribute of the first user and the second user;

training, using the first action data, the second action data, additional action data, the first process data, the second process data, additional process data, the process similarity data, additional process similarity data, the user characteristic data, and additional user characteristic data, a predictive model that is configured to classify actions as performing a process that is similar or not similar to the first process and the second process;

receiving data that identifies actions performed by a particular user;

applying the predictive model to one or more of the actions initiated by the particular user;

based on applying the predictive model to the one or more of the actions initiated by the particular user, classifying a process performed by the one or more of the actions as similar to the first process and the second process; and

based on classifying the process initiated by the one or more of the actions as similar to the first process and the second process, assigning the common demographic, geographic, or employment attribute to the particular user.

2. The method of claim 1, comprising:

receiving data confirming that the particular user possesses the common demographic, geographic, or employment attribute; and

updating the predictive model based on the data confirming that the particular user possesses the common demographic, geographic, or employment attribute.

3. The method of claim 1, comprising:

receiving data confirming that the particular user does not possess the common demographic, geographic, or employment attribute; and

updating the predictive model based on the data confirming that the particular user does not possess the common demographic, geographic, or employment attribute.

4. The method of claim 1, wherein receiving first action data that identifies one or more first actions that were initiated by a first user and second action data that identifies one or more second actions that were initiated by a second user comprises:

receiving first screen capture data from a device performing the one or more first actions;

receiving second screen capture data from a device performing the one or more second actions;

generating the first action data by performing computer vision techniques on the first screen capture data; and

generating the second action data by performing computer vision techniques on the second screen capture data.

5. The method of claim 1, wherein receiving first action data that identifies one or more first actions that were initiated by a first user and second action data that identifies one or more second actions that were initiated by a second user comprises:

receiving first user input data from a device performing the one or more first actions;

receiving second user input data from a device performing the one or more second actions;

generating the first action data by analyzing the first user input data; and

generating the second action data by analyzing the second user input data.

6. The method of claim 1, wherein receiving first process first action data that identifies one or more first actions that were initiated by a first user and second action data that identifies one or more second actions that were initiated by a second user comprises:

receiving first network traffic data from a device performing the one or more first actions;

receiving second network traffic data from a device performing the one or more second actions;

generating the first action data by analyzing the first network traffic data; and

generating the second action data by analyzing the second network traffic data.

7. The method of claim 1, wherein the first process is similar to the second process based on the first process and the second process using a same hardware or a same software.

8. The method of claim 1, wherein the common demographic, geographic, or employment attribute identifies a company, a geographic location, an operational unit, an operational role, or an expertise level.

9. The method of claim 1, wherein an action of the one or more first actions, the one or more second actions, or the actions comprises a key press, a mouse click, a screen touch, a foreground process change, a scene change, a network request, or a network receipt.

10. The method of claim 1, wherein the process, the first process, or the second process comprises viewing a website, receiving a type of email attachment, entering a type of data, performing a web search, or using a type of software.

11. A computer-implemented method comprising:

receiving first action data that identifies one or more first actions and second action data that identifies one or more second actions;

receiving first process data that identifies a first process performed by the one or more first actions and second process data that identifies a second process performed by the one or more second actions;

receiving process similarity data that indicates that the first process is similar to the second process;

receiving first variable action data that identifies a first given action of the one or more first actions that processes first variable information and second variable action data that identifies a second given action of the one or more second actions that processes second variable information;

training, using the first action data, the second action data, additional action data, the first process data, the second process data, additional process data, the process similarity data, additional process similarity data, the first variable action data, the second variable action data, and additional variable action data, a predictive model that is configured to (i) classify one or more actions as performing a process that is similar or not similar to the first process and the second process and (ii) identify a given action among the one or more actions that processes variable information;

receiving data that identifies actions;

applying the predictive model to one or more of the actions;

based on applying the predictive model to the one or more of the actions, (i) classifying a process performed by the one or more of the actions as similar to the first process and the second process and (ii) identifying a given action of the one or more of the actions that processes variable information; and

based on classifying the process as performed by the one or more of the actions as similar to the first process and the second process and on identifying the given action of the one or more of the actions that processes variable information, providing, for output, (i) data identifying the one or more of the actions as similar to the first process and the second process and (ii) data identifying the given action as processing variable information.

12. The method of claim 11, comprising:

receiving data confirming that the process is similar to the first process and the second process; and

updating the predictive model based on the data confirming that the process is similar to the first process and the second process.

13. The method of claim 11, comprising:

receiving data confirming that the process is not similar to the first process and the second process; and

updating the predictive model based on the data confirming that the process is not similar to the first process and the second process.

14. The method of claim 11, wherein receiving first action data that identifies one or more first actions and second action data that identifies one or more second actions comprises:

receiving first screen capture data from a first device performing the one or more first actions;

receiving second screen capture data from a second device performing the one or more second actions;

generating the first action data by performing computer vision techniques on the first screen capture data; and

generating the second action data by performing computer vision techniques on the second screen capture data.

15. The method of claim 11, wherein receiving first action data that identifies one or more first actions and second action data that identifies one or more second actions comprises:

receiving first user input data from a first device performing the one or more first actions;

receiving second user input data from a second device performing the one or more second actions;

generating the first action data by analyzing the first user input data; and

generating the second action data by analyzing the second user input data.

16. The method of claim 11, wherein receiving first action data that identifies one or more first actions and second action data that identifies one or more second actions comprises:

receiving first network traffic data from a first device performing the one or more first actions;

receiving second network traffic data from a second device performing the one or more second actions;

generating the first action data by analyzing the first network traffic data; and

generating the second action data by analyzing the second network traffic data.

17. The method of claim 11, wherein an action of the one or more first actions, the one or more second actions, or the actions comprises a key press, a mouse click, a screen touch, a foreground process change, a scene change, a network request, or a network receipt.

18. The method of claim 11, comprising:

based on applying the predictive model to the one or more of the actions, identifying an additional action of the one or more of the actions that is a superfluous action for performance of the process performed by the one or more of the actions.

19. The method of claim 11, comprising:

determining a type of variable corresponding to the given action; and

providing, for output, the type of variable corresponding to the given action.

20. The method of claim 11, wherein the data identifying the given action as performing the variable information comprises a name for the variable information based on the given action.