Semantic data generation

• US 9,959,317 B2
• Filed: 03/05/2014
• Issued: 05/01/2018
• Est. Priority Date: 03/05/2014
• Status: Active Grant

• Current Assignee: Streamline Licensing LLC (IP Edge LLC)
• Original Assignee: Empire Technology Development LLC (Allied Inventors Management, LLC)
• Inventors: Fang, Jun
• Primary Examiner(s): To, Baoquoc N
• Application Number:
  US14/434,227
• Publication Number:
  US 20160253323A1
• Time in Patent Office: 1,518 Days
• Field of Search: 707748
• CPC Class Codes: G06F 16/2453 Query optimisation
• CPC Class Codes: G06F 16/24575 using context
• CPC Class Codes: G06F 16/24578 using ranking
• CPC Class Codes: G06F 40/30 Semantic analysis
• CPC Class Codes: G06N 5/04 Inference or reasoning models

Abstract

In some examples, a computing device may be configured to simulate the deduction process of human mind by generating new data based on existing data and newly received data that is semantically relevant to the existing data.

Claims

1. A method performed by a computing device, which is communicatively coupled to a first database and a second database, for semantic data projection in an artificial intelligence system, the method comprising:

selecting, from the first database, a first subset of one or more first semantic formulae based on a comparison of respective weights of each of the one or more first semantic formulae and a predetermined threshold weight value, wherein the respective weights of each of the one or more first semantic formulae are assigned to the first subset of one or more first semantic formulae based on a particular topical term;

selecting, from the second database, a second subset of one or more second semantic formulae based on the particular topical term, wherein the second subset of the one or more second semantic formulae is semantically relevant to the first subset of the one or more first semantic formulae;

generating one or more third semantic formulae based on the first subset of the one or more first semantic formulae and the second subset of the one or more second semantic formulae, wherein the generating the one or more third semantic formulae includes generating one or more conclusions based on the first subset of the one or more first semantic formulae and the second subset of the one or more second semantic formulae, thereby enabling simulation of human deduction process in the artificial intelligence system; and

reverting to the selection of the first subset of the one or more first semantic formulae when one or more conditions are not met, wherein the one or more conditions include, at least, a duration of the semantic data projection, a count of times of the reverting, or a count of the generated one or more third semantic formulae.

2. The method of claim 1, wherein each of the one or more weights indicates an importance level of the corresponding first semantic formula with respect to a topic.

3. The method of claim 1, wherein the selecting the first subset of the one or more first semantic formulae includes:

determining the particular topical term;

selecting one of the one or more weights that corresponds to the particular topical term and is greater than the predetermined threshold weight value; and

selecting one of the one or more first semantic formulae that corresponds to the selected weight as one of the first subset of the one or more first semantic formulae.

4. The method of claim 1, further comprising:

converting the first subset of the one or more first semantic formulae into one or more first standard semantic formulae; and

converting the second subset of the one or more second semantic formulae into one or more second standard semantic formulae.

5. The method of claim 4, wherein the generating comprises:

identifying a portion of irresolvable semantic data in one of the one or more first standard semantic formulae;

identifying a negative form of the portion of irresolvable semantic data in one of the one or more second standard semantic formulae;

removing the portion of irresolvable semantic data from the one of the one or more first standard semantic formulae;

removing the negative form of the portion of irresolvable semantic data from the one of the one or more second standard semantic formulae; and

semantically combining a remaining portion of the one of the one or more first standard semantic formulae and a remaining portion of the one of the one or more second standard semantic formulae.

6. The method of claim 1, further comprising including the generated one or more third semantic formulae in the first database.

7. A system, comprising:

a database identifier, implemented at least in part in hardware, configured to;

identify a first database that includes one or more first semantic formulae, each of which is assigned with one or more weights based on a particular topical term, and

identify a second database that includes one or more second semantic formulae;

a data selector, implemented at least in part in hardware and communicatively coupled to the database identifier, configured to select a first subset of the one or more first semantic formulae based on a comparison of the one or more weights and a predetermined threshold weight value;

a semantic relevancy determination module, implemented at least in part in hardware and communicatively coupled to the data selector, configured to identify, from the second database, a second subset of the one or more second semantic formulae based on the particular topical term, wherein the identified second subset of the one or more second semantic formulae is semantically relevant to the first subset of the one or more first semantic formulae;

a semantic formula generator, implemented at least in part in hardware and communicatively coupled to the semantic relevancy determination module, configured to;

generate one or more third semantic formulae based on the first subset of the one or more first semantic formulae and the second subset of the one or more second semantic formulae, wherein the generation of the one or more third semantic formulae includes generation of one or more conclusions based on the first subset of the one or more first semantic formulae and the second subset of the one or more second semantic formulae, so as to enable simulation of human deduction process in the system; and

a terminator, implemented at least in part in hardware and communicatively coupled to the data selector, configured to terminate the system when at least one or more conditions are met, wherein the one or more conditions include, at least, duration of a semantic data projection, a count of times of reverting to the selection of the first subset of the one or more first semantic formulae, or a count of the generated one or more third semantic formulae.

8. The system of claim 7, wherein each of the one or more weights indicates an importance level of the corresponding first semantic formula with respect to a topic.

9. The system of claim 7, wherein the data selector is further configured to:

determine the particular topical term;

identify one of the one or more weights that corresponds to the particular topical term and is greater than the predetermined threshold weight value; and

identify one of the one or more first semantic formulae that corresponds to the identified weight as one of the first subset of the one or more first semantic formulae.

10. The system of claim 7, further comprising a data convertor configured to:

convert the first subset of the one or more first semantic formulae into one or more first standard semantic formulae; and

convert the second subset of the one or more second semantic formulae into one or more second standard semantic formulae.

11. The system of claim 10, wherein the semantic formula generator is further configured to:

identify a portion of irresolvable semantic data in one of the one or more first standard semantic formulae;

identify a negative form of the portion of irresolvable semantic data in one of the one or more second standard semantic formulae;

remove the portion of irresolvable semantic data from the one of the one or more first standard semantic formulae;

remove the negative form of the portion of irresolvable semantic data from the one of the one or more second standard semantic formulae; and

semantically combine a remaining portion of the one of the one or more first standard semantic formulae and a remaining portion of the one of the one or more second standard semantic formulae.

12. The system of claim 7, wherein the semantic formula generator is further configured to include the generated one or more third semantic formulae in the first database.

13. A non-transitory computer-readable medium that stores executable-instructions that, when executed, cause one or more processors of a computing device, which is communicatively coupled to a first database and a second database, to perform operations comprising:

identifying the first database that includes one or more first semantic formulae, each of which is assigned with one or more weights;

selecting a first subset of the one or more first semantic formulae based on a comparison of the one or more weights and a predetermined threshold weight value;

converting the first subset of the one or more first semantic formulae into one or more first standard semantic formulae;

identifying the second database that includes one or more second semantic formulae;

identifying, from the second database, a second subset of the one or more second semantic formulae that is semantically relevant to the first subset of the one or more first semantic formulae;

converting the second subset of the one or more second semantic formulae into one or more second standard semantic formulae, wherein the identifying the second subset of the one or more second semantic formulae includes;

identifying a portion of irresolvable semantic data in one of the one or more first standard semantic formulae; and

identifying a negative form of the portion of irresolvable semantic data in one of the one or more second standard semantic formulae; and

generating one or more third semantic formulae based on the one or more first standard semantic formulae and the one or more second standard semantic formulae, wherein the generating the one or more third semantic formulae includes;

removing the portion of irresolvable semantic data from the one of the one or more first standard semantic formulae;

removing the negative form of the portion of irresolvable semantic data from the one of the one or more second standard semantic formulae; and

semantically combining a remaining portion of the one of the one or more first standard semantic formulae and a remaining portion of the one of the one or more second standard semantic formulae to generate the one or more third semantic formulae, thereby enabling simulation of human deduction process.

14. The computer-readable medium of claim 13, wherein each of the one or more weights indicates an importance level of the corresponding first semantic formula with respect to a topic.

15. The computer-readable medium of claim 13, wherein the selecting includes:

determining a first topical term;

identifying one of the one or more weights that corresponds to the first topical term and is greater than the predetermined threshold weight value; and

identifying one of the one or more first semantic formulae that corresponds to the identified weight as one of the first subset of the one or more first semantic formulae.

16. The computer-readable medium of claim 13, further comprises reverting to the selection when one or more conditions are not met, wherein the one or more conditions include, at least, a duration of a semantic data projection, a count of times of the reverting, or a count of the generated one or more third semantic formulae.

17. The computer-readable medium of claim 13, wherein the second database includes more semantic formulae than the first database.

18. The computer-readable medium of claim 13, wherein the first database includes data previously stored on the computing device and the second database includes data instantaneously collected by an input device.