KR-20260063070-A - question and answer system

Abstract

The present invention relates to a question and answer system. The question-and-answer system according to the present invention can vectorize question information input through an input/output device, then extract a question/answer text module with high cosine similarity to the vectorized question information from a database, and then transmit the extracted question/answer text module as question/answer information to the input/output device. According to the present invention, the search speed can be increased by extracting a question/response text module with high cosine similarity to a vectorized question text, and a question/response text module with high reliability can be extracted for the question text.

Inventors

• 안성민
• 박동길

Assignees

• 주식회사 에이아이오투오

Dates

Publication Date

20260507

Application Date

20241030

Claims (3)

1. A database storing one or more chunked question/response text modules; An input/output device capable of receiving query information and outputting query/response information; and A question response system characterized by comprising: a question response server that vectorizes the question information input through the above input/output device, extracts the question/response text module with high cosine similarity to the vectorized question information from the above database, and transmits the extracted question/response text module as the question/response information to the above input/output device.
2. In Article 1, The query information input from the above input/output device is It may be query text or query voice information, The above question and answer server is A question and answer system characterized by extracting a question text from the question voice information input from the input/output device, vectorizing the extracted question text and storing it in the database, extracting the question/answer text module with high cosine similarity to the vectorized question text, converting it into question/answer information, and transmitting it to the input/output device.
3. In paragraph 2, The above question and answer server is Includes a text spelling check and correction module, A question response system characterized by the text spelling check and correction module checking for spelling errors in the query text of the query information received from the input/output device, automatically correcting the query text with spelling errors if there are spelling errors in the query text, and then vectorizing the corrected query text.

Description

Question and answer system The present invention relates to a question and answer system, and more specifically, to a question and answer system capable of vectorizing question information input through an input/output device, extracting a question/answer text module with high cosine similarity to the vectorized question information from a database, and then using the extracted question/answer text module as question/answer information to output question/answer information with high similarity to the question information through an input/output device. A chatbot is a technology that applies conversational artificial intelligence (AI) technologies, such as Natural Language Processing (NLP), to understand user questions and automatically respond to them. Recently, chatbots for task-oriented dialogue are required to create value in the industry. Purpose-driven conversations refer to questions related to industries and businesses, such as hotel reservations or restaurant orders. Because these conversations vary depending on the unique data, operational manuals, or business knowledge of each industry and company, it can be difficult for generative chatbots to respond automatically. Accordingly, there is a demand for the ability to create chatbots tailored to each industry or company, or to automate tasks while understanding user questions. To meet these requirements, chatbot builders are provided that can understand natural language as much as possible and automatically determine which flow to use to automate tasks according to predefined flows set by individual company chatbot managers. However, to automatically classify flows, machine learning techniques such as supervised learning can be used, and for this, training example data is required. These machine learning-based chatbot services also have several problems. First, it requires investment in cost and time to build query example data. Generating large-scale query example candidate data demands significant time and expense. Secondly, there is the difficulty of establishing pre-flows and conversational intent. It can be difficult to predict whether a defined flow matches a user query. As a result, chatbot managers must create unnecessary flows in advance or perform redundant tasks. In addition, operational and monitoring costs may be incurred. Separate monitoring personnel are required to maintain the quality of the chatbot service. This results in additional operational costs, and monitoring personnel must perform tasks such as verifying new user queries and generating new flows and example queries/answers. This increases the workload of chatbot managers and becomes a factor in raising costs. As such, efficiently managing and operating chatbot services is a complex task, and automated methods are required to overcome these problems. In addition, as Large Language Models (LLMs) have made innovative advancements in the field of natural language processing, they are being utilized in various applications such as chatbot translators and summaries. In addition, there is a demand for conversational systems that can overcome the shortcomings of large-scale language models in chatbots, understand the user's intent regarding questions entered by the user, and generate answers with high reliability or accuracy. FIG. 1 is a block diagram illustrating a question-and-answer system according to an embodiment of the present invention. FIG. 2 is a flowchart illustrating a question-and-answer system according to an embodiment of the present invention. The object of the present invention is to provide a question and answer system capable of improving the search speed for a question/answer text module corresponding to query information and improving the reliability or accuracy of the extracted question/answer text module, and specific advantages and novel features will become more apparent from the following detailed description and preferred embodiments associated with the accompanying drawings. In addition, it should be noted that in assigning reference numbers to the components of each drawing in this specification, identical components are given the same number as much as possible, even if they are shown in different drawings. Furthermore, the present invention may be implemented in various different forms and is not limited to the embodiments described herein. Also, in describing the present invention, detailed descriptions of related known technologies that could unnecessarily obscure the essence of the invention are omitted. A preferred embodiment of the title according to the present invention will be described below with reference to the drawings. As illustrated in FIGS. 1 and 2, an embodiment of a question-and-answer system according to the present invention is illustrated. FIG. 1 is a block diagram illustrating a question-and-answer system according to one embodiment of the present invention. As illustrated in these drawings, the question and answer system accordi