Dialogue System Database Selection for Year-Specific Answers
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing dialogue systems using generative AI face inaccuracies in answering product-related inquiries due to variations in product specifications over time, despite the use of Retrieval Augmented Generation (RAG) technologies.
Innovation Solution
A dialogue system that utilizes a document database switching mechanism to identify the appropriate database for searching based on user queries, and generates prompts for generative AI using search results, ensuring accurate answers by vectorizing documents and utilizing CRM data when necessary.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If RAG (Retrieval Augmented Generation) is used to combine generative AI with search system, then the system can generate answers reflecting specialized knowledge and latest knowledge, but answers to inquiries about products may still be inaccurate when product specifications differ by year of manufacture
Solution Approach 1:
The patent segments the database system into multiple separate databases organized by product category and time period (year of manufacture). Each database contains specifications for products within a specific category and year range, allowing the system to precisely target the relevant database based on the inquiry's product and time context, thereby improving answer accuracy while maintaining manageable database complexity through structured organization.
Solution Approach 2:
The patent implements preliminary action by pre-organizing and storing product specification data in multiple databases according to product category and year of manufacture before any user inquiry occurs. This pre-structured organization enables the system to quickly identify and retrieve the correct specifications without needing to process or filter through all available data during the actual inquiry, thus improving response accuracy and efficiency.
2Device complexity
If a single database is used to store all product information, then the system structure is simple, but the system cannot accurately distinguish between products of the same name from different years with different specifications
Solution Approach 1:
The patent divides a single comprehensive database into multiple segmented databases, each dedicated to a specific product category and time period. This segmentation allows the system to maintain relatively simple individual database structures while achieving high precision in product specification retrieval by selecting the appropriate segmented database based on the product category and manufacturing year mentioned in the user's inquiry.
Solution Approach 2:
The patent applies local quality by creating databases with specialized structures optimized for specific product categories and time periods rather than using a uniform structure for all products. Each database is tailored to contain only the relevant specifications for its designated category and year range, enabling precise retrieval while keeping each individual database manageable in size and complexity.
Data Source
AI summary
A dialogue system is provided with a designator configured to designate a database to be used in a search from a plurality of databases on the basis of a use's query, a generator configured to generate a prompt to be input to a generative AI on the basis of a search result of a search using the designated database and the query, and an outputter configured to output an output of the generative AI in response to the prompt as an answer to the query.