Language-Model Document Search Using Similar Text Hash Tags
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing document search systems require significant user effort and time due to reliance on input search words, which may not accurately reflect the user's intended search intent.
Innovation Solution
A document search device and method utilizing a language model to generate similar texts and search hash tags based on user input, enabling efficient document retrieval by matching these tags with document-specific hash tags, and outputting relevant documents.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a user inputs search words directly to search for documents, then the search process is simple to operate, but the search accuracy deteriorates because the search words may not accurately reflect the user's intended search intent
Solution Approach 1:
The patent introduces an intermediary system consisting of a language model and hash tag generation mechanism. The user's input is transformed into multiple similar texts by the language model, which are then converted into hash tags for document retrieval. This intermediary process bridges the gap between simple user input and accurate document retrieval, maintaining both ease of operation and search accuracy.
Solution Approach 2:
The patent segments the search process into distinct stages: input reception, similar text generation through language model, hash tag creation, and document retrieval. By dividing the search process into these modular segments, the system can apply specialized processing at each stage, improving overall accuracy while keeping the user interface simple.
2Measurement precision
If multiple processing steps are added to improve search accuracy, then the search accuracy improves, but the time and labor required for document search increases
Solution Approach 1:
The patent performs preliminary actions by pre-processing user input into multiple similar texts and generating hash tags before the actual document retrieval. The language model generates similar texts in advance, and hash tags are created beforehand, so that when documents need to be retrieved, the system can quickly match pre-generated hash tags against document indexes, reducing real-time processing time.
Solution Approach 2:
The patent replaces manual, time-consuming search processes with automated language model processing and algorithmic hash tag generation. Instead of manually crafting multiple search queries or manually reviewing documents, the system automatically generates similar texts, converts them to hash tags, and retrieves documents through efficient hashing operations, significantly reducing time and labor.
3Measurement precision
If the system generates multiple similar texts and hash tags automatically, then the search accuracy improves, but the device complexity increases
Solution Approach 1:
The patent implements self-service by enabling the system to automatically generate similar texts and hash tags without requiring external intervention or complex configuration. The language model autonomously processes user input and generates appropriate variations, while the hash tag generation mechanism automatically converts these texts into searchable tags, reducing the need for manual system management despite the increased processing capabilities.
Data Source
AI summary
A document search device includes a memory storing instructions; and one or more processors configured to execute the instructions to: receive a prompt from a user to generate a similar text similar to a search text for document search, generate, based on the prompt, the similar text using a language model by machine learning, generate a search hash tag for the document search based on the search text and the similar text, search for a document based on the search hash tag, and output the searched document.