KR102573140B1 - Ai-based acceptance prediction ststem - Google Patents

Abstract

A method for evaluating an AI-based applicant's possibility of passing a target school, performed by a computing device, is disclosed.
According to an embodiment, a passability evaluation method includes receiving first data corresponding to a result of an applicant's entrance exam, receiving second data corresponding to the applicant's school record, and based on the second data generating at least one first sequence used in a first process of assigning a weight to a concentration region of the first data, the same number of first sequences for the first data as the number of the first sequences. generating two sequences, mapping the at least one first sequence to the second sequence one-to-one, and generating a third sequence for each of the mapped first and second sequences, Applying a first process for , applying a second process to the first data based on the third sequence and the second sequence, and applying the second process to the first data and pre-learned data. Evaluating the applicant's chances of passing based on an artificial neural network may be included.

Description

AI-based acceptance prediction system {AI-BASED ACCEPTANCE PREDICTION STSTEM}

The following examples relate to an AI-based pass prediction system, and in more detail, using the information related to the entrance examination of existing applicants who have pass or failure information of the target school (international school, university, graduate school) that wants to go to school and the information of the test subject It relates to a pass prediction method and device suitable for predicting the pass or fail of a target school.

The number of students who want to study abroad is increasing every year as the domestic entrance exam becomes more intense and the interest in global talents demanded by the future society increases. However, as going abroad requires a lot of money and effort, you have no choice but to be very careful in making this decision. Therefore, it is necessary to set a goal for a school with a high possibility of admission according to the different characteristics of each individual and start preparing after accurately recognizing the needs of each school.

The problem to be solved by the present invention is to provide a more accurate and suitable pass prediction system for each applicant situation by reducing accuracy errors and deviations from the methodological limitations of traditional statistical theory or probability-based prediction systems during overseas admission screening.

An AI-based method for evaluating the likelihood of an applicant's acceptance into a target school, performed by a computing device, comprising: receiving first data corresponding to a result of an entrance examination of the applicant; receiving second data corresponding to the applicant's school records; generating at least one first sequence used in a first process of assigning a weight to a concentration region of the first data, based on the second data; generating the same number of second sequences for the first data as the number of the first sequences; The first process is applied to the first data by mapping the at least one first sequence to the second sequence one-to-one and generating a third sequence for each of the mapped first and second sequences. doing; applying a second process to the first data based on the third sequence and the second sequence; and evaluating a possibility of passing the applicant based on the first data to which the second process is applied and the previously trained artificial neural network.

The generating of the first sequence may include determining information on the number of the first sequence based on information on the number of items included in the report; determining information on the number of individual elements included in each of the first sequences based on information on the number of sections of an entrance examination; and generating the at least one first sequence based on the first sequence number information and the individual element number information.

The generating of the first sequence may further include determining a value of an individual element included in the first sequence based on score information for each item included in the report, wherein the individual element is the first sequence. It is determined as a single value for each sequence, and the step of applying the first process generates the third sequence based on multiplication between the mapped first sequence and corresponding individual elements in the second sequence, The applying of the second process may include generating a fourth sequence based on a sum operation of values of corresponding individual elements among the generated third sequences; and generating at least one fifth sequence based on a multiplication operation for values of individual elements corresponding to the fourth sequence and each of the at least one second sequence.

The target schools include any one of international schools, American private schools, boarding schools, overseas law schools, overseas law schools, overseas dental schools, overseas dental schools, overseas pharmacy schools, and overseas pharmacy schools, and the entrance exam is, It may include at least one of SAT, ACT, AP, SAT2, IB, ALevel, TOEFL, IELTS, GCSE, MYP, AMC, MUN, GED, SSAT, ISEE, MAP, LSAT, MCAT, DAT, and PCAT.

A computing device that performs an AI-based applicant's passability evaluation method for a target school includes a communication unit; and a processor, wherein the processor receives first data corresponding to a result of the applicant's entrance examination, receives second data corresponding to the applicant's school record score, and based on the second data, At least one first sequence used in a first process of assigning a weight to a concentration region of the first data is generated, and second sequences of the first data having the same number as the number of the first sequences are generated. and mapping the at least one first sequence to the second sequence one-to-one, and generating a third sequence for each of the mapped first sequence and second sequence, thereby performing a first process for the first data. , and applies a second process to the first data based on the third sequence and the second sequence, and applies the second process to the first data to which the second process is applied and based on the pre-learned artificial neural network of the applicant. Evaluate your chances of success.

In the present invention, it is possible to more accurately predict whether an applicant will pass or fail to apply for a school by using AI. More specifically, unlike the conventional method of predicting pass or fail based on a simple entrance test and statistics on the numerical value of GPA, the correlation between the scores for each element of the entrance exam and the scores for each element included in the GPA. There is an advantage in providing a predictive model with improved accuracy compared to the conventional method through a means for predicting pass using an artificial neural network pre-learned to consider degrees.

1 is a diagram showing the overall configuration of an AI-based pass prediction system according to the present invention.
2 is a configuration diagram of a computing device for predicting an applicant's acceptance probability according to an embodiment of the present invention.
3 is a diagram for schematically explaining an AI-based passability evaluation method according to the present invention.
4 is a flowchart for explaining a pass rate evaluation method according to the present invention.
5 is a block diagram for explaining the structure of a computing device according to an embodiment.

Specific structural or functional descriptions of the embodiments according to the concept of the present invention disclosed in this specification are only illustrated for the purpose of explaining the embodiments according to the concept of the present invention, and the embodiments according to the concept of the present invention These may be embodied in various forms and are not limited to the embodiments described herein.

Embodiments according to the concept of the present invention can apply various changes and can have various forms, so the embodiments are illustrated in the drawings and described in detail herein. However, this is not intended to limit the embodiments according to the concept of the present invention to specific disclosures, and includes modifications, equivalents, or substitutes included in the spirit and scope of the present invention.

Terms such as first or second may be used to describe various components, but the components should not be limited by the terms. The above terms are used only for the purpose of distinguishing one component from another component, for example, without departing from the scope of rights according to the concept of the present invention, a first component may be named a second component, Similarly, the second component may also be referred to as the first component.

It is understood that when an element is referred to as being "connected" or "connected" to another element, it may be directly connected or connected to the other element, but other elements may exist in the middle. It should be. On the other hand, when an element is referred to as “directly connected” or “directly connected” to another element, it should be understood that no other element exists in the middle. Expressions describing the relationship between components, such as "between" and "directly between" or "directly adjacent to" should be interpreted similarly.

Terms used in this specification are only used to describe specific embodiments, and are not intended to limit the present invention. Singular expressions include plural expressions unless the context clearly dictates otherwise. In this specification, terms such as "comprise" or "having" are intended to designate that a feature, number, step, operation, component, part, or combination thereof exists, but one or more other features or numbers, It should be understood that the presence or addition of steps, operations, components, parts, or combinations thereof is not precluded.

Unless defined otherwise, all terms used herein, including technical or scientific terms, have the same meaning as commonly understood by one of ordinary skill in the art to which the present invention belongs. Terms such as those defined in commonly used dictionaries should be interpreted as having a meaning consistent with the meaning in the context of the related art, and unless explicitly defined in this specification, it should not be interpreted in an ideal or excessively formal meaning. don't

The present invention predicts the probability of an applicant's acceptance using an artificial neural network trained on the basis of information on existing applicants who have a history of applying for a target school where the applicant wishes to go to school. Through accurate probability prediction, it helps applicants learn.

Hereinafter, preferred embodiments of the present invention will be described in detail with reference to the accompanying drawings.

1 is a diagram showing the overall configuration of an AI-based pass prediction system according to the present invention.

The pass prediction system may include a user terminal device 110, a network 120, and a computing device 130 corresponding to a server that performs a pass prediction process.

The user terminal device 100 is a digital device equipped with a memory means such as a desktop computer, a notebook computer, a workstation, a PDA, a webpad, a mobile phone, etc. that can be connected to the computing device 130 and equipped with a microprocessor and having arithmetic capability. means

The network 120 may be a high-speed backbone network of a large-scale communication network capable of transmitting and receiving data, and Wi-Fi, WiGig, and Wireless Broadband for providing wired and wireless Internet or high-speed multimedia services. It may be a next-generation wireless network including Internet, Wibro), World Interoperability for Microwave Access (Wimax), and the like.

According to an embodiment of the present invention, the computing device 130 receives information about applicants through the network 120, analyzes the received information, and provides information related to the probability of acceptance to the university to which you apply to the user terminal device 100. It refers to a device that provides, and such a computing device may be configured in the form shown in FIG. 2, for example.

2 is a configuration diagram of a computing device for predicting an applicant's acceptance probability according to an embodiment of the present invention.

Referring to FIG. 2, the computing device 130 is largely divided and may include an interface unit 210, a service execution module 220, a database module 230, and the like, and the service execution module 220 is an information management unit 221 ), information acquisition unit 222, item classification unit 223, modeling unit 224, AI engine 225, consulting unit 226, etc., and the database module 230 is a member information DB (231), an existing applicant DB (232), and various test and admission screening DB (233).

Referring to FIG. 2 , the information management unit 221 may provide a function of registering as a user member by storing related member information in the member information DB 231 when a member registration request is received from each user through the network 120. there is. In addition, it is possible to perform login of members accessing the server through the member information DB 232 .

Next, the information acquisition unit 222 searches the member information DB 232 or requests the test subject to obtain information about the test subject when requesting a pass prediction service for a school desired by the applicant corresponding to the test subject. .

The item classification unit 223 classifies applicant information and test subject information according to the examination items for each school.

The modeling unit 224 is an element that generates an AI engine that predicts an applicant's acceptance probability using an arbitrary artificial neural network model, and may generate an artificial neural network pre-learned through training data.

In one embodiment, the modeling unit 224 creates an AI engine by learning the information of existing applicants stored in the existing applicant DB 232 (entrance exam score information, school report score information, personal statement, volunteer activity details, etc.). The modeling unit 224 may generate various AI-based pass prediction models and may further include a separate verification unit (not shown) that performs verification of these models.

The AI engine unit 225 inputs and analyzes applicant information acquired through the information acquisition unit 222 into the AI-based pass prediction model generated by the modeling unit 224 .

The consulting unit 226 provides the probability of passing as a numerical value based on the results analyzed by the AI engine unit 225. Here, when predicting the probability of acceptance to the target school, classification accuracy is calculated. The classification accuracy is the probability of predicting that a person who applied to the target school and failed to fail and the person who applied to the target school and passed was predicted to pass. It means the average value by summing the probabilities.

In addition, the consulting unit 226 refers to the admission process of the target school selected by the test target in various tests and admission screening DB 233 to create preparations or supplements necessary to increase the likelihood of passing the test target, or to prepare for the target school. A function to generate explanatory data can be provided.

Finally, the database module 230 includes information about students' entrance examinations and case study information optimized over several years. Information on the top 10-30, top 30-50 universities in the US, top 3 universities in Canada, top 10 universities in the UK, Hong Kong, Singapore, and emerging prestigious schools, as well as contests, internships, volunteer activities, and academics. It includes statistics on performance trends and expected passing rates by university, by official score, region, and entrance examination. The part described above in relation to the criteria stored in the database module 230 is only exemplary, and the criterion for university selection, the type of information, etc. may not be limited to the examples presented above.

In one embodiment of the present invention, when an applicant accesses the computing device 130 corresponding to the pass prediction service server through log-in, the information management unit 221 in the service execution module 220 selects a school, a pass prediction request item, etc. A screen including is retrieved and transmitted to the user terminal 110 of the corresponding applicant through the interface unit 210 . At this time, the applicant selects the target school to receive the pass prediction service using the school selection item on the main screen, and requests the pass prediction service for the target school he or she wishes to advance to by clicking the pass rate prediction request item. can

When an applicant requests a pass rate prediction service, the information obtaining unit 222 acquires prediction target information of the corresponding applicant through a search of the member information DB 232 and transfers it to the item classification unit 223 .

The item classification unit 223 classifies the applicant information according to the evaluation items for each school.

The schools that applicants wish to enter may include international schools, American private schools, boarding schools, overseas law schools, overseas law schools, overseas dental schools, overseas dental schools, overseas pharmacy schools, and overseas pharmacy schools. There are so many different entrance exam requirements that it is difficult to prepare for them.

Essays, official certification scores, and portfolios can be included in the evaluation items for each school. In particular, it is essential to submit an essay when studying in the United States. Essay topics vary from school to school, but generally include reasons for choosing to study abroad and reasons for applying to the school. In addition, there are cases where you have to write an article based on the individual values that schools value, and it is important to accurately analyze it because the standards are very different for each school.

The official certification score corresponding to the entrance test score also varies greatly from school to school and from country to country. These include the SAT, ACT, AP, IB, LNAT, LSAT, MCAT, DAT, and PCAT. It is necessary to understand the pros and cons of each official test and prepare for admission through customized analysis at the right time. For example, both the SAT and ACT are American college entrance qualification tests, but unlike the SAT, the ACT has a separate science subject, and the ACT math can be used with a calculator. You need to check if it suits you. AP and IB are prerequisites for university courses in the US and UK, respectively. LNAT is for admission to British law school, LSAT is for admission to US law school, MCAT is for admission to US medical school, DAT is for admission to US dental school, and PCAT is for admission to US law school. is a required test for admission to the University of Pharmacy in the United States.

The portfolio may include research experiences, volunteer activities, internship experiences, completed courses, club activities, and awards. Portfolios have different requirements depending on majors or colleges and are a part that can make use of individual characteristics, so preparing for each school and individual is the most important thing.

An AI-based evaluation model of the computing device 130 is created by acquiring information on existing applicants from the existing applicant DB 232 . The AI-based evaluation model may be implemented based on various artificial neural networks, such as an ensemble model, a logistic regression model, a neural network model, a Bayesian classification model, and a support vector machine model.

Even after selecting a model, it is necessary to look at information aspects, diagnose problems, and perform tuning to improve the model. Tuning involves changing the architecture or tweaking hyperparameters. Such work requires inferring the cause of the problem and finding a way to improve the expected phenomenon.

The computing device 130 according to the present invention evaluates the applicant's chances of passing by using an artificial neural network, which is an AI model pre-learned through entrance exam results and school records, among the above-described entrance exam methods, thereby making it more effective than general statistical methods. It can produce precise prediction results. The computing device may provide more precise pass prediction information through additional analysis of the portfolio information and essay information described above in addition to the prediction result using the artificial neural network.

3 is a diagram for schematically explaining an AI-based passability evaluation method according to the present invention.

Referring to FIG. 3 , first data 301 corresponding to an admission test result input to a computing device performing an AI-based passability evaluation method is subjected to a first process 310 to emphasize information related to passability evaluation. and a second process 320 may be applied. Here, the first process 310 and the second process 320 are operations performed to emphasize some of the input information to highlight main information, which may be referred to as attention.

The first data 301 may be an entrance exam score, for example, an SAT score, and may be data composed of scores for each subject in the form of a sequence. Those skilled in the art will understand that the first data 310 may correspond to an arbitrary entrance test other than the SAT depending on the applicable university.

The computing device generates a first sequence 312 generated based on the applicant's report card 311 corresponding to the first data 301 and the first data 101 through the convolution module 313. The first process 310 may be applied to the first data 101 by generating the third sequence 316 through a channelwise product 315 of the second sequence 314 .

The computing device generates a fourth sequence 322 in the form of a single sequence by performing a sum operation 321 on values included in all the generated third sequences 316, and the fourth sequence 322 and the second sequence 322 are combined. The second process 320 may be applied by generating the fifth sequence 324 through the channel-by-channel multiplication operation 323 between the sequences 314 . For example, when there are 12 third sequences 316, the value of the generated fourth sequence 322 may be determined based on a summation result of values included in each of the 12 third sequences 316. there is. The computing device generates the same fourth sequence 322 as many as the second sequence 314 and performs a channel-by-channel multiplication operation between the generated fourth sequences and the second sequence 314, so that the fifth sequence ( 324) can be created. A specific method of performing the first process 310 and the second process 320 will be described in more detail with reference to the accompanying drawings.

The computing device may evaluate the applicant's passability based on the first data 301 by inputting the fifth sequence 324 to the pre-learned artificial neural network 330 . The first process 310 and the second process 320 are applied, and the first data 301 in which predetermined information is emphasized can be evaluated for the applicant's passability through the pre-learned artificial neural network 330 . The artificial neural network 330 may be an artificial neural network pre-learned through learning data generated based on existing entrance exam result data. More specifically, the learning data for training the artificial neural network 330 is the first process 310 and the second process 320 described above for each of the big data (previous applicant's entrance exam scores) corresponding to the existing entrance exam results. ) can be created by applying The artificial neural network 330 may be trained in advance to output a passing probability (probability) corresponding to the learning data based on the above learning data.

4 is a flowchart for explaining a pass rate evaluation method according to the present invention.

Referring to FIG. 4 , the pass rate evaluation method according to the present invention includes, first, a computing device receiving first data about an applicant's entrance exam result (S101). For example, the first data may be score information corresponding to a result of an entrance exam, in which the entrance exam includes subjects A, subjects B, and subjects C, and the applicant receives 80 points in subject A, 90 points in subject B, and 75 points in subject C. When a point is obtained, the first data may have a shape of [80, 90, 75]. 80, 90, and 75 presented in the examples may be referred to as "individual elements" throughout this specification.

In addition, the pass rate evaluation method according to the present invention may include a step of receiving, by a computing device, second data corresponding to the applicant's GPA (S102).

In addition to the entrance exam, the grade report may include information about the grades of the tests the applicant has taken while performing school life. For example, the grades may be data in the form of accumulated grades for each subject (Korean language, mathematics, science, etc.), and depending on the implementation example, arbitrary elements such as evaluation scores for usual behavior and evaluation scores for attendance It will be appreciated by those skilled in the art.

An example of the grade in the report may correspond to Table 1.

subject sexual Korean (1st grade) Average 85 Mathematics (Year 1) average 88 Science (Year 1) Average 83 ?? ?? English (Grade 3) Average 90 Social Studies (Grade 3) Average 84 PE (Grade 3) Average 90

Referring to Table 1, the grades may have a form in which the average score for each grade and subject is mapped to the corresponding subject, but this is an example, and the type of subject included in the grade is not limited thereto, but is arbitrary. It will be appreciated by those skilled in the art that the evaluation elements of can be scored and included.

In the above situation, the form of the second data may be [85, 88, 83, ??, 90, 84, 90], which is a form of listing grades from Korean (1st grade) to PE (3rd grade). 85, 88, 83, 90, 40, and 90 presented in the examples may be referred to as “individual elements” throughout this specification.

As will be described later, the computing device determines the number of first sequences (for example, 7) according to the number of items included in the school report, and determines the number of items (subjects) included in the entrance exam result and the school report. Individual elements constituting the first sequence may be determined according to the grades.

The contents of the internal report, the method for determining the number of first sequences, and the method for determining the value corresponding to the first sequence, which are exemplarily presented through Table 1, are only exemplary, and an embodiment of the present invention is shown in Table 1. It will be appreciated by those skilled in the art that it is not limited to the examples presented.

Next, the pass rate evaluation method according to the present invention includes generating, by a computing device, at least one first sequence for applying a first process to an entrance exam result based on the school report (S200).

The computing device may determine information on the number of first sequences to be generated later based on the items included in the report card. For example, if grades for 10 subjects are included in the grade, the computing device may generate 10 first sequences. In another embodiment, when 10 subjects included in the grade report are divided into 3 different categories, the computing device may generate 3 first sequences. For example, subjects included in school grades may be classified into different categories based on Principal Component Analysis (PCA), and the number of first sequences may be determined according to the number of categories.

The computing device may determine information on the number of individual elements to be included in each first sequence based on the number of items included in the first data. More specifically, the computing device may determine the number of individual elements included in the first sequence to correspond to the number of areas (subjects) included in the first data, the admission test result.

The computing device may generate the first sequence based on the determined number information of the first sequence and the number of individual elements constituting the first sequence. A value of an individual element included in each first sequence may be determined as a single value for each first sequence. In addition, the determined single value may be determined based on the score of an internal report card corresponding to the first sequence. For example, the values of individual elements included in the first sequence generated corresponding to the second subject (mathematics (first grade)) of the grade report may be determined based on the average score of the second subject. More specifically, when the second subject is "mathematics (1st grade)" and the average score is 0 to 30, the value of the corresponding individual element is 1, 31 to 60 points, the value is 3, 61 to 80 points , If the value of the corresponding individual element is 5, and the score is 81 to 100, if the value of the corresponding individual element is determined to be 7, and if the applicant scored an average of 88 points for 1st grade mathematics, the second subject corresponds to the second subject. Each of the individual elements included in 1 sequence may be determined to be 7.

The computing device may further include generating second sequences equal to the number of generated first sequences ( S300 ). For example, when 5 first sequences are generated, the computing device may generate 5 second sequences for admission test results using a convolution module. The second sequence may refer to a feature sequence in which a weight is assigned to second data corresponding to an admission test result through a convolution module.

In the passability evaluation method, the computing device maps the first sequence to the second sequence one-to-one, and generates a third sequence for each of the mapped first and second sequences, so that the first sequence and the second sequence are mapped to the first sequence and the second sequence respectively. 1 process may be applied (S400). The computing device may apply the first process to the input test result by generating the third sequence based on Equations 1 and 2.

is your grade, Is a sequence generated through the values of individual elements determined based on the total number (t) determined based on the subject of the grade and the average score for each subject, is the first sequence, Is A second sequence generated based on the first data, which is the result of the entrance exam, through the convolution module by the number of represents a third sequence, The operation is a multiplication operation for each channel, and t means the number of subjects included in the second data. For reference, each first sequence The number of individual elements constituting may be u corresponding to the number of subjects in the first data.

Specifically, the computing device maps the first sequence and second sequence By multiplying the values of individual elements corresponding to , the third sequence can create For example, the mapped first sequence and second sequence The product of the values of each ith individual element is a third sequence By assigning to the i-th individual element of, the mapped first sequence and second sequence A third sequence corresponding to this can be created.

The passability evaluation method may include applying, by the computing device, a second process to an entrance exam result based on the generated third and second sequences ( S500 ). The computing device may apply the second process to the admission test result based on Equations 3 to 5.

is the fourth sequence, is the ith individual element of the fourth sequence, Is the i-th individual element of the k-th third sequence, is an extended fourth sequence extended by generating as many fourth sequences as the number of second sequences; is the k-th fourth sequence of the extended fourth sequences; represents a fifth sequence.

Specifically, the computing device generates all third sequences A fourth sequence that is a single sequence based on the sum of each corresponding individual element of can create The computing device is the fourth sequence cast generated by the number (t) of and the second sequence is mapped 1 to 1, and the mapped and the second sequence The fifth sequence multiplication of the values of the corresponding individual elements of The fifth sequence by assigning to the corresponding individual element of can create In this way, the computing device is mapped to all and the second sequence for By generating, it is possible to apply the second process to the first data.

The passability evaluation method may include evaluating, by a computing device, the passability of an applicant who has provided the first data based on the first data to which the second process is applied and the pretrained artificial neural network 330 (S600). For example, the computing device may evaluate the applicant's passability using the pre-learned artificial neural network 330 based on the first data (admission test result) to which the first process and the second process are applied. For example, the evaluation result may be provided in the form of an acceptance probability of the applicant school.

According to an embodiment, the computing device may calculate a gradient for the passability evaluation result and provide information on an influence of the first data or the second data on the evaluation result based on the calculated gradient. More specifically, the computing device calculates a first gradient based on the evaluation result for each individual element of the first data (the grade of each subject included in the entrance exam result), and the individual element of the second data (included in the grade) The second gradient may be calculated based on the grades of the individual subjects), and information on the influence of each individual element on the evaluation result may be provided based on the calculated first gradient and the second gradient. For example, the computing device may provide a list in which individual elements having a high pass prediction change rate are calculated by rank according to changes in scores of individual elements based on the calculated gradient. The above list can be used as a basis for establishing a college application strategy in consideration of the average test score of the year in the admissions consulting process, and can be used as a basis for establishing a college admissions strategy for the next year.

5 is a block diagram for explaining the structure of a computing device according to an embodiment.

Referring to FIG. 5 , a computing device 500 according to an embodiment includes a processor 520 . The computing device 500 may further include a memory 510 and a communication interface 530 . The processor 510, the memory 530 and the communication interface 550 may communicate with each other through a communication bus (not shown).

The processor 520 receives first data corresponding to the result of the applicant's entrance exam, receives second data corresponding to the applicant's grade point average, and determines a concentration area of the first data based on the second data. generating at least one first sequence used in a first process of assigning weights, generating a second sequence for the first data having the same number as the number of first sequences, and generating at least one first sequence; By mapping the second sequence one-to-one and generating a third sequence for each of the mapped first and second sequences, the first process for the first data is applied, and the third sequence and the second sequence are Based on this, a second process may be applied to the first data, and the likelihood of passing the applicant may be evaluated based on the first data to which the second process is applied and the previously trained artificial neural network.

Memory 510 may be volatile memory or non-volatile memory.

In addition, the processor 520 may execute a program and control the computing device 500 . Program codes executed by the processor 520 may be stored in the memory 510 . The computing device 500 may be connected to an external device (eg, a personal computer or network) through an input/output device (not shown) and exchange data. The computing device 500 may be mounted on a server. Those skilled in the art will understand that the computing device 500 can perform the method described above with reference to FIGS. 1 to 4 .

The embodiments described above may be implemented as hardware components, software components, and/or a combination of hardware components and software components. For example, the devices, methods and components described in the embodiments may include, for example, a processor, a controller, an arithmetic logic unit (ALU), a digital signal processor, a microcomputer, a field programmable gate (FPGA). array), programmable logic units (PLUs), microprocessors, or any other device capable of executing and responding to instructions. A processing device may run an operating system (OS) and one or more software applications running on the operating system. A processing device may also access, store, manipulate, process, and generate data in response to execution of software. For convenience of understanding, there are cases in which one processing device is used, but those skilled in the art will understand that the processing device includes a plurality of processing elements and/or a plurality of types of processing elements. It can be seen that it can include. For example, a processing device may include a plurality of processors or a processor and a controller. Other processing configurations are also possible, such as parallel processors.

Software may include a computer program, code, instructions, or a combination of one or more of the foregoing, which configures a processing device to operate as desired or processes independently or collectively. The device can be commanded. Software and/or data may be any tangible machine, component, physical device, virtual equipment, computer storage medium or device, intended to be interpreted by or provide instructions or data to a processing device. , or may be permanently or temporarily embodied in a transmitted signal wave. Software may be distributed on networked computer systems and stored or executed in a distributed manner. Software and data may be stored on one or more computer readable media.

The method according to the embodiment may be implemented in the form of program instructions that can be executed through various computer means and recorded on a computer readable medium. The computer readable medium may include program instructions, data files, data structures, etc. alone or in combination. Program commands recorded on the medium may be specially designed and configured for the embodiment or may be known and usable to those skilled in computer software. Examples of computer-readable recording media include magnetic media such as hard disks, floppy disks and magnetic tapes, optical media such as CD-ROMs and DVDs, and magnetic media such as floptical disks. - includes hardware devices specially configured to store and execute program instructions, such as magneto-optical media, and ROM, RAM, flash memory, and the like. Examples of program instructions include high-level language codes that can be executed by a computer using an interpreter, as well as machine language codes such as those produced by a compiler.

As described above, although the embodiments have been described with limited drawings, those skilled in the art can apply various technical modifications and variations based on the above. For example, the described techniques may be performed in an order different from the method described, and/or components of the described system, structure, device, circuit, etc. may be combined or combined in a different form than the method described, or other components may be used. Or even if it is replaced or substituted by equivalents, appropriate results can be achieved.

Therefore, other implementations, other embodiments, and equivalents of the claims are within the scope of the following claims.

Claims (7)

In the method of evaluating the likelihood of passing an AI-based applicant to a target school, performed by a computing device,
receiving first data corresponding to a result of an applicant's entrance examination;
receiving second data corresponding to the applicant's school records;
generating at least one first sequence used in a first process of assigning a weight to a concentration region of the first data, based on the second data;
generating the same number of second sequences for the first data as the number of the first sequences;
The first process is applied to the first data by mapping the at least one first sequence to the second sequence one-to-one and generating a third sequence for each of the mapped first and second sequences. doing;
applying a second process to the first data based on the third sequence and the second sequence; and
Evaluating the applicant's passability based on the first data to which the second process is applied and a previously trained artificial neural network;
Generating the first sequence,
determining information on the number of the first sequence based on information on the number of items included in the report;
determining information on the number of individual elements included in the first sequence based on information on the number of sections of an entrance examination; and
generating the at least one first sequence based on the first sequence number information and the individual element number information;
Determining a value of an individual element included in the first sequence based on score information for each item included in the report
Including more,
The individual element is determined as a single value for each of the first sequences,
Applying the first process,
generating the third sequence based on multiplication between the mapped first sequence and corresponding individual elements in the second sequence;
Applying the second process,
generating a fourth sequence based on a sum operation of values of corresponding individual elements among the generated third sequences;
generating at least one fifth sequence based on a multiplication operation for values of individual elements corresponding to the fourth sequence and each of the at least one second sequence;
The target school is
Including any one of international schools, American private schools, boarding schools, overseas law schools, overseas law schools, overseas dental schools, overseas dental schools, overseas pharmacy universities, and overseas pharmacy graduate schools;
The entrance exam is
At least one of SAT, ACT, AP, SAT2, IB, ALevel, TOEFL, IELTS, GCSE, MYP, AMC, MUN, GED, SSAT, ISEE, MAP, LSAT, MCAT, DAT, PCAT,
Generating the fifth sequence,
Determine the fifth sequence based on Equations 1 to 5,
Equation 1 above is
, ego,
Equation 2 above is
, ego,
Equation 3 above is
ego,
Equation 4 above is
, ego,
Equation 5 above is
ego,
remind is your grade, above Is a sequence generated through the values of individual elements determined based on the total number (t) determined based on the subject of the grade and the average score for each subject, the above is the first sequence, the said A second sequence generated based on the first data, which is the result of the entrance exam, through a convolution module by the number of represents a third sequence, The operation is a multiplication operation for each channel, where t is the number of subjects included in the second data, and u is each first sequence is the number of subjects in the first data as the number of individual elements constituting
remind Is the fourth sequence, the Is the ith individual element of the fourth sequence, Is the i th individual element of the k th third sequence, wherein Is an extended fourth sequence extended by generating as many fourth sequences as the number of second sequences, Is the k-th fourth sequence of the extended fourth sequence, the Is the fifth sequence,
an passability evaluation method that calculates a gradient for a passability evaluation result and further provides information on an influence of the first data or the second data on the evaluation result based on the calculated gradient;
A computing device capable of calculating a gradient for a passability evaluation result and providing information on an influence of the first data or the second data on the evaluation result based on the calculated gradient. delete delete delete delete delete delete