Document-Grounded Language Processing for Reliable Response Generation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The question response apparatus disclosed in Patent Literature 1 does not verify the correctness of responses from external search engines, potentially leading to incorrect answers.
Innovation Solution
An information processing apparatus that acquires a target text, extracts relevant documents, rewrites the text using these documents, generates a corresponding text with a machine learning model, and outputs the result with document information to enhance reliability.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If a question response apparatus uses an external search engine to generate responses, then the apparatus can provide answers to questions, but the reliability of the responses cannot be verified and may be incorrect
Solution Approach 1:
The patent introduces a language model as an intermediary component between the external search engine and the final response output. The language model processes the extracted entities and search results, generating verified responses that maintain reliability while preserving productivity. This intermediary layer ensures that responses are not directly taken from potentially unreliable external sources but are instead generated through a controlled linguistic processing pipeline.
Solution Approach 2:
The system implements a feedback mechanism where the language model's generation process is guided by verified information from external searches. The model receives feedback in the form of extracted entities and search results, which constrain and inform its response generation, ensuring that the output is both productive and reliable.
2Ease of operation
If a language model generates text based on input text, then language processing can be performed, but the reliability of the processing result cannot be ensured
Solution Approach 1:
The patent applies preliminary action by extracting entities from the input text before the language model generates its response. This pre-processing step identifies key information that must be contained in the final response, ensuring that the language model operates with predetermined constraints that guarantee reliability. The entity extraction acts as a preliminary verification step that guides subsequent generation.
Solution Approach 2:
The extracted entities serve as feedback to the language model, guiding its text generation process. The model receives these entities as input constraints and must incorporate them into its output, creating a feedback loop that ensures the processing result is both linguistically natural and factually accurate according to the extracted key information.
Data Source
AI summary
To improve reliability of a result of language processing carried out with use of a machine learning model, an information processing apparatus includes at least one processor that carries out: an acquisition process of acquiring a target text; an extraction process of extracting a document related to the target text; a rewriting process of rewriting the target text with use of the document; a generation process of generating a text corresponding to the rewritten target text with use of a machine learning model trained to generate a text based on an input text; and an output process of outputting a result obtained by adding information identifying the document to the generated text.