KR-20260067486-A - Multi-turn execution method to minimize information loss and processing volume for LLM chatbot development

Abstract

A multi-turn execution method for minimizing information loss and throughput for LLM chatbot development is disclosed. The multi-turn execution method comprises the following steps: when the LLM chatbot provides a "first response" to a user's "first question," summarizing the "first question" and the "first response" (S10); storing the newly summarized "query information" and completing preparations to respond to the "1+n-th question" (S20); the LLM chatbot providing the "1+n-th response" using the "query information" and the "1+n-th question" (S30); summarizing "new query information" using the "1+n-th question," the "1+n-th response," and the "query information" (S40); overwriting the "query information" with the newly summarized "new query information" and completing preparations to respond to the next new question (S50). and includes a step (S60) of repeating the steps S30, S40, and S50 above.

Inventors

• 전채남
• 이태환
• 손형진
• 손상우
• 지동주
• 송은지

Assignees

• (주) 더아이엠씨

Dates

Publication Date

20260513

Application Date

20241105

Claims (2)

1. When the LLM chatbot provides a "first response" to the user's "first question," a step (S10) of summarizing the "first question" and the "first response"; A step (S20) of storing the newly summarized "query information" and completing preparations to respond to the "1+nth question"; A step (S30) in which the LLM chatbot provides a "1+nth response" using the above "query information" and "1+nth question"; A step (S40) of summarizing "new query information" using "1+nth question", "1+nth response", and the above "query information"; The newly summarized "new query information" is overwritten over the "query information" and the step of completing preparations to respond to the new next question (S50); and A multi-turn execution method comprising a step (S60) of repeating the steps S30, S40, and S50 above.
2. In paragraph 1, The above steps S10 and S20 are the first turn (S100), and A multi-turn execution method in which the above S30, S40, and S50 steps are the 1+nth turn (S200).

Description

Multi-turn execution method to minimize information loss and processing volume for LLM chatbot development The present invention relates to a multi-turn execution method for natural conversation with a chatbot, and more specifically, to a multi-turn execution method capable of minimizing the amount of information loss and information processing volume when developing a chatbot. Recently, with the development of artificial intelligence technology, the fields providing chatbot services are increasing. While search engines provide desired results immediately with a single input, chatbots utilizing generative AI pursue multi-turn conversations capable of continuous dialogue rather than short, one-word conversations in order to seek natural conversations similar to human interaction. In order to increase user satisfaction with chatbot services and ensure continuous service, it must be possible to accurately understand user queries and provide appropriate answers during continuous conversations with people (multi-turn conversations). When implementing multi-turn functionality using such generative AI, one must be careful that the chatbot understands the flow of the conversation and maintains the context appropriately, while minimizing the amount of information processed regarding past conversations. In order to implement the correct multi-turn function, past conversation information must be used as the basis for answering new questions, so how past information is remembered and processed is key to implementing this function. However, due to the opportunity cost between maintaining flow and context and minimizing information processing volume, a process for retaining appropriate information is currently required. FIG. 1 is a drawing for explaining a multi-turn execution method according to an embodiment of the present invention. FIG. 2 is a flowchart illustrating a multi-turn execution method according to an embodiment of the present invention. The present invention is susceptible to various modifications and may take various forms, and embodiments are to be described in detail in the text. However, this is not intended to limit the invention to the specific disclosed forms, and it should be understood that the invention includes all modifications, equivalents, and substitutions that fall within the spirit and scope of the invention. Similar reference numerals have been used for similar components in the description of each figure. Terms such as "first," "second," etc., may be used to describe various components, but said components should not be limited by said terms. The above terms are used solely for the purpose of distinguishing one component from another. The terms used in this application are used merely to describe specific embodiments and are not intended to limit the invention. The singular expression includes the plural expression unless the context clearly indicates otherwise. In this application, terms such as "comprising" or "consisting of" are intended to specify the existence of the features, numbers, steps, actions, components, parts, or combinations thereof described in the specification, and should be understood as not precluding the existence or addition of one or more other features, numbers, steps, actions, components, parts, or combinations thereof. Unless otherwise defined, all terms used herein, including technical or scientific terms, have the same meaning as generally understood by those skilled in the art to which the present invention pertains. Terms such as those defined in commonly used dictionaries should be interpreted as having a meaning consistent with their meaning in the context of the relevant technology, and should not be interpreted in an ideal or overly formal sense unless explicitly defined in this application. Terms such as "first" or "second" may be used to describe various components, but said components should not be limited by said terms. For the sole purpose of distinguishing one component from another, for example, without departing from the scope of rights according to the concept of the present invention, the first component may be named the second component, and similarly, the second component may be named the first component. When it is stated that one component is "connected" or "connected" to another component, it should be understood that while it may be directly connected or connected to that other component, there may also be other components in between. Conversely, when it is stated that one component is "directly connected" or "directly connected" to another component, it should be understood that there are no other components in between. Other expressions describing the relationship between components, such as "between" and "exactly between," or "adjacent to" and "directly adjacent to," should be interpreted in the same way. Hereinafter, preferred embodiments of the present invention will be described in more detail with reference to the drawings. FIG. 1