SYSTEMS AND METHODS FOR CREATING AN ARTIFICIAL INTELLIGENCE

• US 20150046376A1
• Filed: 08/06/2013
• Published: 02/12/2015
• Est. Priority Date: 08/06/2013
• Status: Abandoned Application

• Current Assignee: Ilya Geller
• Original Assignee: Ilya Geller
• Inventors: Geller, Ilya
• Application Number:
  US13/960,397
• Publication Number:
  US 20150046376A1
• Time in Patent Office: Days
• US Class Current: 706/12
• CPC Class Codes: G06N 5/02 Knowledge representation; Symbolic representation

Abstract

A computer system implemented method of creating and using artificial intelligence wherein the system includes a first table including at least one textual portion and at least one context phrase contained in the textual portion and the computer system is capable of communication via a network with the user using at least one electronic device, the method comprising: receiving at least one textual input from the user, extracting at least one portion of the textual input from the user and at least one context phrase therefrom, comparing each portion extracted from the textual input from the user to other portions extracted from the textual input from the user according to a first matching algorithm that utilizes the context phrases of each respective portion, and storing in the first table, the portions and respective context phrases that were extracted from the textual input from the user that satisfy the matching algorithm.

Claims

1. A computer system implemented method of creating and using an artificial intelligence (AI) clone comprising the following operations of a computer system:

for said AI clone, receiving text into the computer system from one or more sources;

for said AI clone, obtaining respective paragraphs from the text received for the AI clone;

at least some of the paragraphs comprising multiple sentences, and at least some of the sentences comprising multiple clauses identified based upon figures of speech and punctuation;

obtaining a first set of context phrases from the received paragraphs, which context phrases are obtained from the clauses and are indicative of the context of the paragraphs;

obtaining respective weights of the context phrases using parameters related to frequency of occurrence of a context phrase relative to other context phrases or to absolute number of occurrences of a context phrase therein;

storing the context phrases and the paragraphs as structured data in one or more tables to thereby create an initial version of the AI clone;

improving the initial version of the AI clone by adding paragraphs and a second set of context phrases from text subsequently supplied to the computer system by the source of the text that was used to create the initial version of the AI clone and from one or more other sources, including one or more instructors, and by selectively deleting data from the one or more tables, to thereby create an improved AI clone; and

using the improved AI clone and selectively using the initial version of the AI clone to (a) answer questions posed by users through a process comprising using a compatibility test matching context phrases related to the respective questions to context phrases related to the AI clone through a compatibility algorithm relating weights of context phrases related to a question and weights of context phrases related to the AI clone, and (b) to direct advertisements to the AI clone, using for the purpose a matching algorithm that uses selected matching criteria in comparing context phrases obtained from the questions or advertisements with said structured data in said one or more tables, which matching algorithm relates context phrases related to advertisements to context phrases related to the AI clone and takes into account respective weights of the context phrases that the matching algorithm relates;

wherein the AI clone is configured to replace a human source of information in answering a user'"'"'s question and assist advertisers in selecting an AI clone that is likely to be receptive to a selected advertisement to thereby direct the advertisement to the At clone based on the content of the question and the advertisement.

2. The method of claim 1 further including adding, to the AI clone to which an advertisement has been directed in said using step, one or more paragraphs of that advertisement and one or more context phrases obtained from the last-recited paragraphs, to thereby further improve the AI clone.

3. The method of claim 1 wherein each context phrase is a combination of a noun with other parts of speech at least one of which is a verb or an adjective.

4. The method of claim 1 in which the step of improving the AI clone further comprises directing a question to an AI clone and, if an answer from the AI clone is not satisfactory to an instructor, storing a correct answer in the AI clone and deleting from the AI clone at least some of the history leading to the unsatisfactory answer.

5. The method of claim 1 in which the compatibility algorithm takes into account the frequency of occurrence of context phrases obtained from the question and context phrases obtained from a paragraph of the AI clone against which paragraph the question is being matched in said compatibility test.

6. The method of claim 1 in which the matching algorithm takes into account the frequency of occurrence of context phrases obtained from the AI clone and context phrases obtained from an advertisement.

7. The method of claim 1 in which the compatibility algorithm takes into account sizes of the paragraphs from which the context phrases are obtained.

8. The method of claim 1 in which said weights comprise a metric related to frequency of occurrence of context phrases in the respective paragraphs from which the context phrases were obtained.

9. The method of claim 1 in which said weights comprise a metric related to frequency of occurrence of context phrases relative to the frequency of occurrence of other context phases in the paragraphs from which they were obtained.

10. A non-transitory computer readable storage medium containing a program in a non-transitory form that when executed in a computer system causes the system to create and use an artificial intelligence (AI) clone of respective entity in a computer-implemented process comprising:

for the AI clone, receiving respective text into the computer system from one or more sources;

for the AI clone, obtaining respective paragraphs from the text received for the AI clone;

at least some of the paragraphs comprising multiple sentences, and at least some of the sentences comprising multiple clauses identified based upon figures of speech and punctuation;

obtaining a first set of respective context phrases from the received paragraphs, which context phrases are obtained from the respective clauses and are indicative of the context of the respective paragraphs;

obtaining respective weights of the context phrases using parameters related to frequency of occurrence of a context phrase relative to other context phrases or to absolute number of occurrences of a context phrase therein;

storing the context phrases and the paragraphs as structured data in one or more tables to thereby create an initial AI clone;

improving the AI clone by adding paragraphs and a second set of context phrases from text subsequently supplied to the computer system by the source of the text that was used to create the initial AI clone and from one or more other sources, including one or more instructors, and by selectively deleting data from the one or more tables, to thereby create an improved AI clone; and

using the improved AI to (a) answer questions posed by users through a process comprising using a compatibility test matching context phrases related to the respective questions to context phrases related to the AI clone through a compatibility algorithm relating weights of context phrases related to a question and weights of context phrases related to the AI clone, and (b) direct advertisements to AI clone, using for the purpose a matching algorithm that uses selected matching criteria in comparing context phrases obtained from the questions or advertisements with said structured data in said one or more tables, which matching algorithm relates context phrases related to advertisements to context phrases related to the AI clone and takes into account respective weights of the context phrases that the matching algorithm relates;

wherein the AI clone is configured to replace human sources of information in answering a user'"'"'s question and assist advertisers in selecting an AI clone that is likely to be receptive to an advertisement to thereby direct the advertisement to an AI clone based on the content of the question and the advertisement.

11. The storage medium of claim 10 in which said program is further configured to cause the computer system to add, to the AI clone to which an advertisement has been directed in said using step, one or more paragraphs of that advertisement and one or more context phrases obtained from the last-recited paragraphs, to thereby further improve the AI clone.

12. The storage medium of claim 10 in which said program is further configured to cause the computer system to cause each context phrase to be a combination of a noun with other parts of speech at least one of which is a verb or an adjective.

13. The storage medium of claim 10 in which said program is further configured to cause the computer system to improve the AI clone by directing a question to the AI clone and, if an answer from the AI clone is not satisfactory to an instructor, to store a correct answer in the AI clone and delete from the AI clone at least some of the history leading to the unsatisfactory answer.

14. The storage medium of claim 10 in which said program is further configured to cause the computer system to take into account in the compatibility algorithm the frequency of occurrence of context phrases obtained from the question and context phrases obtained from a paragraph of the AI clone against which paragraph the question is being matched in said compatibility test.

15. The storage medium of claim 10 in which said program is further configured to cause the computer system to take into account in the matching algorithm the frequency of occurrence of context phrases obtained from the AI clone and context phrases obtained from an advertisement.

16. The storage medium of claim 10 in which said program is further configured to cause the computer system to take into account, in said compatibility algorithm, sizes of the paragraphs from which the context phrases are obtained.

17. The storage medium of claim 10 in which said program is further configured to cause the computer system to calculate said weights as a metric related to frequency of occurrence of context phrases in the respective paragraphs from which the context phrases were obtained.

18. The storage medium of claim 10 in which said program is further configured to cause the computer system to calculate said weights as a metric related to frequency of occurrence of context phrases relative to the frequency of occurrence of other context phases in the paragraphs from which they were obtained.