Historical Q&A Search Using Topic Aggregation and Sub-Dimension Tags
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The inefficiency of searching through large quantities of historical dialog content in intelligent dialog systems, where content is typically shown sequentially, reduces the searching efficiency and accuracy.
Innovation Solution
An information searching method that aggregates historical question and answer information into topics, using sub-dimension tags to refine granularity and facilitate flexible searching, improving efficiency and accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If historical dialog contents are shown in sequential order of dialog times, then the complete historical content is preserved, but the content searching efficiency is reduced
Solution Approach 1:
The patent segments historical dialog contents into multiple topics based on semantic analysis and aggregation. Each topic represents a cluster of related dialog rounds, allowing users to navigate through organized categories rather than sequential entries. This segmentation enables rapid location of target content by selecting relevant topics instead of scrolling through all historical dialogs in time order.
Solution Approach 2:
The patent introduces a new organizational dimension for historical content by aggregating dialog rounds into topics based on semantic similarity and content relationships. This transforms the single-dimensional time-based sequence into a multi-dimensional structure where content can be accessed both by time and by topic category, enabling efficient searching through hierarchical organization.
2Measurement precision
If manual sliding is used to search through historical dialog contents, then the complete content can be reviewed, but the searching accuracy is reduced
Solution Approach 1:
The patent introduces topic aggregations as intermediary structures between the user's search query and the raw historical dialog contents. Topics serve as mediators that pre-organize and categorize content, allowing users to directly access relevant information without manually sliding through all entries. This intermediary layer enhances searching accuracy by filtering and organizing content before presentation to the user.
Solution Approach 2:
The patent replaces the mechanical manual sliding operation with an automated topic-based navigation system. Instead of requiring users to manually scroll through sequential dialog entries, the system automatically aggregates content into topics and presents them in an organized manner, substituting manual mechanical interaction with intelligent automated organization and presentation.
3Quantity of substance
If a large quantity of historical dialog contents is generated, then more comprehensive information is available, but the difficulty of finding target content increases
Solution Approach 1:
The patent merges multiple dialog rounds into unified topics based on semantic analysis and content correlation. By combining related conversations into single topic units, the system reduces the perceived quantity of separate entries while maintaining comprehensive information coverage. This merging strategy makes large volumes of historical content more manageable and easier to navigate by presenting them as organized thematic collections rather than individual scattered entries.
Data Source
AI summary
The present disclosure provides an information searching method and information searching apparatus, a computer device, and a storage medium. The method includes: obtaining and presenting a plurality of question and answer topics in response to a searching operation for historical question and answer information, wherein the question and answer topics are determined based on aggregation results obtained by aggregating a plurality of rounds of historical question and answer information; obtaining, in response to any one of the plurality of question and answer topics represented being triggered, at least one sub-dimension tag under the question and answer topic; and presenting each sub-dimension tag, and presenting a sub-aggregation result under a selected sub-dimension tag according to timing information of historical question and answer information, wherein the sub-aggregation result is obtained by aggregating the aggregation result according to a target dimension indicated by the selected sub-dimension tag.