KR-20260063499-A - METHOD AND SYSTEM FOR RETRIEVING INFORMATION

KR20260063499AKR 20260063499 AKR20260063499 AKR 20260063499AKR-20260063499-A

Abstract

A method and a system for information retrieval are provided. The method comprises the steps of obtaining a query and inputting the query into a previously trained information retrieval model and determining a plurality of search target information corresponding to the query among a plurality of candidate information using the output of the information retrieval model, wherein the plurality of search target information includes a first search target information and a second search target information, wherein the first search target information is automatically determined among the plurality of candidate information using the output of the information retrieval model based on similarity with the query, and the second search target information may be automatically determined among the first remaining plurality of search target information excluding the first search target information using the output of the information retrieval model based on similarity with the first search target information.

Inventors

신재선
김동우
홍혜림
김준이
이민영

Assignees

삼성에스디에스 주식회사

Dates

Publication Date: 20260507
Application Date: 20241030

Claims (9)

In a method performed by a computing system, Step of obtaining a query; and The method includes the step of inputting the above query into a previously trained information retrieval model and determining a plurality of search target information corresponding to the above query among a plurality of candidate information using the output of the information retrieval model, The above plurality of search target information includes first search target information and second search target information, and The above first search target information is automatically determined from among the plurality of candidate information based on similarity with the query using the output of the information search model, and The above second search target information is automatically determined using the output of the information search model from among the first remaining multiple candidate information excluding the first search target information based on similarity with the above first search target information, Information search methods.
In Article 1, The second search target information is determined from among the first remaining plurality of candidate information based on the difference value of similarity with the first search target information obtained by multiplying the similarity with the query by a weight, Information search methods.
In Article 2, It further includes the step of determining the above weight, The step of determining the above weight is: Step of obtaining a validation data set; and The method includes the step of inputting the verification data set into the information retrieval model and determining the weight using the output of the information retrieval model, The above weights are optimized from pre-set initial weights to maximize the performance of the information retrieval model for the above verification data set, Information search methods.
In Article 1, The above plurality of search target information further includes third search target information, and The above third search target information is automatically determined using the output of the information search model from among the remaining multiple candidate information, excluding the first search target information and the second search target information, based on similarity with the query, similarity with the first search target information, and similarity with the second search target information. Information search methods.
In combination with computing devices, Step of obtaining a query; and To execute the step of inputting the above query into a previously trained information retrieval model and determining a plurality of search target information corresponding to the above query among a plurality of candidate information using the output of the information retrieval model, the data is stored on a computer-readable recording medium, The above plurality of search target information includes first search target information and second search target information, and The above first search target information is automatically determined from among the plurality of candidate information based on similarity with the query using the output of the information search model, and The above second search target information is automatically determined using the output of the information search model from among the first remaining multiple candidate information excluding the first search target information based on similarity with the above first search target information, Computer program.
At least one processor; and It includes at least one memory that stores instructions that cause the at least one processor to perform operations when executed by the at least one processor, and The above operations are, Operation to obtain a query; and The method includes inputting the above query into a previously trained information retrieval model and determining a plurality of search target information corresponding to the above query among a plurality of candidate information using the output of the information retrieval model, wherein The above plurality of search target information includes first search target information and second search target information, and The above first search target information is automatically determined from among the plurality of candidate information based on similarity with the query using the output of the information search model, and The above second search target information is automatically determined using the output of the information search model from among the first remaining multiple candidate information excluding the first search target information based on similarity with the above first search target information, Information retrieval system.
In Article 6, The second search target information is determined from among the first remaining plurality of candidate information based on the difference value of similarity with the first search target information obtained by multiplying the similarity with the query by a weight, Information retrieval system.
In Article 7, The above operations are, It further includes an operation to determine the above weight, The operation for determining the above weight is: The operation of acquiring a validation data set; and The method includes inputting the verification data set into the information retrieval model and determining the weight using the output of the information retrieval model, The above weights are optimized from pre-set initial weights to maximize the performance of the information retrieval model for the above verification data set, Information retrieval system.
In Article 6, The above plurality of search target information further includes third search target information, and The above third search target information is automatically determined using the output of the information search model from among the remaining multiple candidate information, excluding the first search target information and the second search target information, based on similarity with the query, similarity with the first search target information, and similarity with the second search target information. Information retrieval system.

Description

Method and System for Retrieving Information The present disclosure relates to an information retrieval method and a system thereof. Specifically, the present disclosure relates to a method for retrieving information corresponding to a query from an information pool and a system for performing the method. In a method for searching for information corresponding to a query, multiple pieces of information with high similarity to the query can be extracted from an information pool. In this case, among the multiple pieces of information included in the information pool, the top-1 information with the highest similarity to the query can be extracted, and among the extracted multiple pieces of information, the remaining information excluding the top-1 information may be data that has high similarity to the top-1 information. For example, when a query for a request to send general emails to employees is input into a search model, the 'General Email Sending API' with the highest similarity to the query can be extracted from the information pool, and the 'Confidential Email Sending API', which has a high probability of having a high similarity value to the query due to its high similarity to the extracted 'General Email Sending API', can also be extracted. Consequently, a problem may arise where the 'Employee Search API,' which has high similarity to the query but low similarity to the extracted 'General Email Sending API,' is pushed down in search priority to a lower rank than the 'Confidential Email Sending API,' resulting in the 'Employee Search API,' which is actually necessary to process the query request, not being extracted. Therefore, a new information retrieval method is required to solve these problems. FIG. 1 is a configuration diagram illustrating an example of a search management system to which an information retrieval system according to one embodiment of the present disclosure can be applied. FIG. 2 is a flowchart illustrating an information retrieval method according to one embodiment of the present disclosure. FIG. 3 is a flowchart illustrating an example of the overall operation of an information retrieval system according to some embodiments of the present disclosure. FIG. 4 is a flowchart illustrating the process of determining the weight of similarity between search target information and candidate information according to some embodiments of the present disclosure. FIG. 5 is a flowchart illustrating an example of a process for optimizing the weight of similarity between search target information and candidate information according to some embodiments of the present disclosure. FIG. 6 is a block diagram illustrating an example of a computing device for carrying out some embodiments of the present disclosure. Preferred embodiments of the present disclosure are described in detail below with reference to the accompanying drawings. The advantages and features of the present disclosure and the methods for achieving them will become clear by referring to the embodiments described below in detail together with the accompanying drawings. However, the present disclosure is not limited to the embodiments described below but may be implemented in various different forms. The embodiments are provided merely to make the present disclosure complete and to fully inform those skilled in the art of the scope of the invention, and the embodiments of the present disclosure are defined only by the scope of the claims. To avoid obscuring the concepts of the present disclosure, known components may be omitted or illustrated in the form of block diagrams focusing on the core functions of each component. Throughout the present disclosure, the same components are described using the same reference numerals, even if they are shown in different drawings. Unless otherwise defined, all terms used herein (including technical and scientific terms) may be used in a meaning commonly understood by those skilled in the art to which this disclosure pertains. Furthermore, terms defined in commonly used dictionaries are not to be interpreted ideally or excessively unless explicitly and specifically defined otherwise. The terms used herein are for describing embodiments and are not intended to limit the invention. In this disclosure, the singular form includes the plural form unless specifically stated otherwise in the text. Furthermore, the terms used in this disclosure are used merely to describe specific embodiments and are not intended to limit the features, components, sequences, etc. described in the specification. Terms such as “comprises” and/or “comprising” used in this disclosure express the existence of the features, components, steps, actions, and/or combinations thereof described in the specification, and do not exclude the existence or addition of one or more other features, components, steps, actions, and/or combinations thereof. In addition, terms such as 1, 2, A, B, (a), (b), etc. used in the following embodiments a