US-12626188-B2 - Artificial intelligence feedback method and artificial intelligence feedback system

US12626188B2US 12626188 B2US12626188 B2US 12626188B2US-12626188-B2

Abstract

An artificial intelligence (AI) feedback method includes: providing input1 and input2 to a script extractor and executor, providing, by the script extractor and executor, an initial condition to an AI, generating, by the AI, a two-dimensional (2D) information table from the initial condition, adding the popularity frequency value or ranking value to corresponding coordinates in the 2D information table and matching knowledge by an interaction between the AI and the script extractor and executor, generating, by the AI, a 2D pseudolinear transformation table from the 2D information table, and performing, by the AI, pseudolinear transformation multiple times to form a 2D unique characteristic table and deriving information of coordinates whose characteristic vector is not changed as a deep truth value.

Inventors

Seung Chan Lee

Assignees

KOREA HYDRO & NUCLEAR POWER CO., LTD.

Dates

Publication Date: 20260512
Application Date: 20210803
Priority Date: 20200812

Claims (1)

1 . An artificial intelligence (AI) feedback method in which each step is performed by an AI feedback system implemented as a computing device, the AI feedback method comprising: generating input 1 selected from document files and picture files and input 2 selected from key words and intensities of outlines; generating an initial condition for ranking values of information related to input 2 from input 1; providing the initial condition to an AI; generating, using the AI, a two-dimensional (2D) information table in the form of a matrix having a key word as a first coordinate axis and key information as a second coordinate axis based on the ranking value derived from the initial condition; adding, using the AI, the ranking values and learning results acquired by the AI to corresponding coordinates of the 2D information table; generating, using the AI, a 2D pseudolinear transformation table in which the ranking values of the 2D information table are digitized, the 2D pseudolinear transformation table having the same dimension as that of the 2D information table and each coordinate of the 2D pseudolinear transformation table being randomly turned on or off; and performing, using the AI, multiple iterations for a pseudolinear transformation of performing an element-wise product between the 2D pseudolinear transformation table and the 2D information table and then updating the 2D information table, to form a 2D unique characteristic table in which each coordinate has a characteristic vector and deriving information of coordinates whose characteristic vector is not changed among the coordinates of the 2D unique characteristic table as a deep truth value.

Description

TECHNICAL FIELD The present invention relates to an artificial intelligence (AI) feedback method and AI feedback system. BACKGROUND ART A learning method of an AI of a related art has many steps and takes a long time depending on accumulated patterns or methods for realizing the accumulated pattern. The learning method of the AI of the related art is a method in which there should be data unilaterally injected from the outside for an AI to build and recognize knowledge, and thereby find a new equilibrium point. Since the learning method of the AI of the related art receives an enormous amount of information and relies on regression analysis and classification through the information, there are problems in which an exponentially large number of calculations and a long time are required as the learning step proceeds. DISCLOSURE Technical Problem An exemplary embodiment is to provide an artificial intelligence (AI) feedback method and an AI feedback system in which intuitive feedback for learning is efficiently achieved. Technical Solution An exemplary embodiment of the present invention provides an artificial intelligence (AI) feedback method in which each step is performed by an AI feedback system implemented as a computing device including: providing input 1 including a basic material for learning and input 2 including core contents for learning to a script extractor and executor; providing, by the script extractor and executor, an initial condition for a popularity frequency value or a ranking value of information related to input 2 to an AI from input 1; generating, by the AI, a two-dimensional (2D) information table in the form of a matrix having a key word and key information based on the popularity frequency value or the ranking value derived from the initial condition as both axes; adding the popularity frequency value or the ranking value to corresponding coordinates of the 2D information table according to an interaction between the AI and the script extractor and executor and matching knowledge acquired by the AI to the corresponding coordinates; generating, by the AI, a 2D pseudolinear transformation table in which the ranking value of the 2D information table is digitized and each coordinate is randomly turned on or off from the 2D information table, the 2D pseudolinear transformation table having the same area as that of the 2D information table; and performing, by the AI, pseudolinear transformations for calculating the 2D pseudolinear transformation table on the 2D information table multiple times to form a 2D unique characteristic table in which each coordinate has a characteristic vector and deriving information of coordinates whose characteristic vector is not changed among the coordinates of the 2D unique characteristic table as a deep truth value. Input 1 may include a plurality of picture files, and input 2 may include an intensity of an outline. Input 1 may include a plurality of document files, and input 2 may include set words. Another exemplary embodiment of the present invention provides an artificial intelligence (AI) feedback system including: a script extractor and executor receiving input 1 including a basic material for learning and input 2 including core content for learning, and generating an initial condition for a popularity frequency value or ranking value of information related to input 2 from input 1; an AI connected to the script extractor and executor, generating a two-dimensional (2D) information table in the form of a matrix having a key word and key information based on the popularity frequency value or the ranking value derived from the initial condition as both axes, adding the popularity frequency value or the ranking value to corresponding coordinates of the 2D information table, and matching the knowledge acquired by the AI to the corresponding coordinates; and an output unit connected to the AI and outputting a deep truth value derived by the AI from the two-dimensional information table. The AI may generate a 2D pseudolinear transformation table in which the ranking value of the 2D information table is digitized and each coordinate is randomly turned on or off from the 2D information table, the 2D pseudolinear transformation table having the same area as that of the 2D information table, perform pseudolinear transformations for calculating the 2D pseudolinear transformation table on the 2D information table multiple times to form a 2D unique characteristic table in which each coordinate has a characteristic vector and derive information of coordinates whose characteristic vector is not changed among the coordinates of the 2D unique characteristic table as a deep truth value. Advantageous Effects According to an exemplary embodiment, the artificial intelligence (AI) feedback method and the AI feedback system in which intuitive feedback for learning is efficiently performed are provided. DESCRIPTION OF THE DRAWINGS FIG. 1 is a flowchart illustrating an artifici