Artificial intelligence and knowledge based automation enhancement

• US 10,289,959 B2
• Filed: 11/12/2015
• Issued: 05/14/2019
• Est. Priority Date: 05/26/2010
• Status: Active Grant

• Current Assignee: Automation Anywhere, Inc.
• Original Assignee: Automation Anywhere, Inc.
• Inventors: Kakhandiki, Abhijit, Gunapu, Sridhar
• Primary Examiner(s): Starks, Wilbert L
• Application Number:
  US14/939,715
• Publication Number:
  US 20160078368A1
• Time in Patent Office: 1,279 Days
• Field of Search: None
• CPC Class Codes: G06N 20/00 Machine learning
• CPC Class Codes: G06N 3/08 Learning methods
• CPC Class Codes: G06N 5/04 Inference or reasoning models

Abstract

This invention generally relates to a process, system and computer code for updating of computer applications based on collecting automation information related to a current application such as processing power, load, footprint, and performance attributes, determining a system automation profile; using an artificial intelligence based modeler for analyzing data, applying the data to an artificial intelligence model for training and predicting performance, adjusting the artificial intelligence model to achieve an updated automation criteria with optimal values, wherein the optimal values provide input to an automation criteria library for storing and updating a prior automation criteria, and exporting the upgraded automation criteria values for incorporation in a computer-to-be-updated, to achieve a reliable automatic update.

Claims

1. A computerized method for updating a remote computer based on artificial intelligence methods comprises the steps of:

A. (a) utilizing one or more of hard-wired circuitry or software instructions for collecting automation information related to a current application by comparing system inputs and outputs to and from a service provider computer and a remote computer as a candidate for updating, respectively, to stored results in the remote computer to optimize automation fit in the remote computer, such as processing power, load, footprint and performance attributes;

(b) locating application controls;

(c) scanning the computer to extract data on processing power, load, footprint, and performance attributes;

(d) using the data in an automation profile generator for determining a system automation profile;

(e) using an artificial intelligence based modeler for analyzing input and output data;

(f) and applying the data to an artificial intelligence model for training and predicting performance, and wherein if the training and predicting function is not satisfied during test, the artificial intelligence model is adjusted and the process of creating a satisfactory performance model is repeated, until the training and predicting function is satisfied;

(g) then merging the artificial intelligence model within a cloud based database;

(h) adjusting the artificial intelligence model to achieve an updated automation criteria with optimal values, wherein the optimal values provide input to an automation criteria library for storing and updating a prior automation criteria; and

(i) utilizing one or more of hard-wired circuitry or software instruction for exporting the upgraded automation criteria values for incorporation in the remote computer to optimize automation fit in the remote computer such as processing power, load, footprint and performance attributes, and wherein one of hard-wired circuitry or software instruction implements a process for;

B. executing (a) one or more of a series of commands, executable as variables by the remote computer'"'"'s application;

(b) one or more tasks assembled into execution files, which validate the tasks and organize nested tasks, including (i) collecting nested task information for each task, (ii) accounting for all dependencies to insure that files, tasks, and environments for operating the remote computer to achieve optimal automation fit such as processing power, load, footprint and performance attributes.

2. The method of claim 1, wherein the artificial intelligence model utilizes one of supervised and unsupervised learning methods, regression analysis, Bayesian statistical models, and clustering algorithms.

3. The method of claim 1, further including comparing system inputs and automation outputs from the service provider computer to stored results to obtain the optimal automation fit for the remote computer under examination, as a candidate for updating.

4. The method of claim 1, further including utilizing one or more of hard-wired circuitry or software instructions for determining system automation profiles for a remote computer'"'"'s processing power, processing load, speed, processor type, manufacture application footprint, characteristics and specific application.

5. The method of claim 1, further including constructing artificial intelligence based learning mechanisms to derive key automation criteria.

6. The method of claim 1, further including combining artificial intelligence based learning on the service provider system, with broader learning from multiple similar systems, to provide recommendations on automation criteria such as optimal automation speed, best automation technique, and recommendations for optimal delays.

7. The method of claim 1, further including creating portable automation criteria libraries, correlated to a system with a specific automation profiles.

8. A non-transitory computer-readable medium for creating adapters that enable updating a remote computer including steps of:

A. (a) collecting automation information related to a current application by comparing system inputs and outputs to and from a service provider computer system and a remote computer as a candidate for updating, respectively, to stored results in the remote computer as a candidate for updating to optimize automation fit in the remote computer such as processing power, load, footprint and performance attributes;

(b) locating application controls;

(c) scanning the computer application to extract data on processing power, load, footprint, and performance attributes;

(d) using the data in an automation profile generator for determining a system automation profile;

(e) using an artificial intelligence based modeler for analyzing input and output data;

(f) and applying the data to an artificial intelligence model for training and predicting performance, and wherein if the training and predicting function is not satisfied during test, the artificial intelligence model is adjusted and the process of creating a satisfactory performance model is repeated, until the training and predicting function is satisfied;

(g) adjusting the artificial intelligence model to achieve an updated automation criteria with optimal values, wherein the optimal values provide input to an automation criteria library for storing and updating a prior automation criteria; and

(h) exporting the upgraded automation criteria values for incorporation in a remote computer to optimize automation fit in the remote computer such as processing power, load, footprint and performance attributes wherein;

B. one or more tasks in the service provider computer have (a) one or more of a series of commands, executable as variables by the remote computer application;

(b) one or more tasks assembled into execution files, which validate the tasks and organize nested tasks, including (i) collecting nested task information for each task, (ii) accounting for all dependencies to insure that files, tasks, and environments for operating the remote computer to optimize automation fit in the remote computer, such as processing power, load, footprint and performance attributes.

9. A computerized method for creating adapters to update a computer-to-be-updated based on artificial intelligence methods comprises the steps of:

(a) utilizing one or more of hard-wired circuitry or software instructions for collecting automation information related to a current computer by comparing system inputs and outputs to and from a service provider computer and a computer-to-be-updated as a candidate for updating, respectively, to stored results in the remote computer to optimize automation fit in the computer-to-be-updated, such as processing power, load, footprint and performance attributes, as a candidate for updating;

(b) locating application controls;

(c) scanning the computer-to-be-updated to extract data on processing power, load, footprint, and performance attributes;

(d) using the data in an automation profile generator for determining a system automation profile;

(e) using an artificial intelligence based modeler for analyzing input and output data;

(f) and applying the data to an artificial intelligence model for training and predicting performance, and wherein if the training and predicting function is not satisfied during test, the artificial intelligence model is adjusted and the process of creating a satisfactory performance model is repeated, until the training and predicting function is satisfied;

(g) then merging the artificial intelligence model within a cloud based database;

(h) adjusting the artificial intelligence model to achieve an updated automation criteria with optimal values, wherein the optimal values provide input to an automation criteria library for storing and updating a prior automation criteria; and

(i) utilizing one or more of hard-wired circuitry or software instruction for exporting the upgraded automation criteria values for incorporation in the computer-to-be-updated to achieve a reliable automatic update to optimize automation fit in the computer-to-be-updated such as processing power, load, footprint and performance attributes incorporation in the computer-to-be-updated to achieve a reliable automatic update.