Trained Search Query Generation for Accurate Answers
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing sentence generation methods using large language models are prone to inaccuracies when input questions are unclear, insufficient in explanation, or redundant, leading to irrelevant reference information and deteriorated answer accuracy.
Innovation Solution
An information processing apparatus and method that utilizes a trained generation model to generate a search query based on an input question, followed by using the detected information to generate an answer, improving accuracy by aligning with the questioner's intent.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If a feature vector is calculated directly from the input question sentence to acquire reference information, then the process is simple and fast, but the accuracy of the answer deteriorates when the question is unclear, insufficient, or redundant
Solution Approach 1:
The patent introduces a search query as an intermediary between the input question and the reference information retrieval. The search query generation unit creates a refined search query based on the input question, which then serves as the basis for calculating the feature vector and retrieving reference information. This intermediary step filters and clarifies the question intent before processing, improving answer accuracy while managing complexity through modular design.
2Reliability
If reference information is acquired based on the input question sentence, then the system can generate answers, but irrelevant reference information is acquired when the question does not reflect the questioner's intention
Solution Approach 1:
The patent performs preliminary action by generating a search query before retrieving reference information. The search query generation unit processes the input question first to create a refined query that better reflects the questioner's intention. This preliminary processing ensures that the subsequent feature vector calculation and reference information retrieval are based on clarified intent, reducing the loss of original meaning and improving relevance.
3Ease of operation
If a large language model is used to generate answers directly from the input question, then the system is simple to operate, but the answer accuracy is affected by the quality of the input question
Solution Approach 1:
The patent implements self-service by enabling the system to automatically refine its own input through search query generation. The search query generation unit autonomously processes the input question to create a more precise query without requiring manual intervention. This self-service mechanism maintains ease of operation for users while internally improving the quality of information processing to enhance answer accuracy.
Data Source
AI summary
Accuracy of an answer to a question is improved. An information processing apparatus includes: a query generation section that uses at least one generation model, which has been generated by being trained by machine learning so as to be able to generate a search query that is in accordance with an input question, to generate a search query that is for retrieving information related to a target question; and an answer generation section that uses information detected by search with use of the search query to generate an answer to the target question. This information processing apparatus makes it easy to generate an answer that is optimized for a specific application.