Topic Recognition Using Knowledge Base Concept Tagging
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Identifying the topic of a collaborative messaging session can be challenging when keywords are absent or ambiguous terms are used, making it difficult for natural language processing systems to infer the conversation topic accurately.
Innovation Solution
A method and system that utilize a knowledge base to associate additional concepts with topics, allowing processors to tag and replace ambiguous terms with main topics, thereby enhancing topic recognition in written conversations by continuously expanding the knowledge base with information from various sources.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional keyword-based topic identification is used, then the system is simple to operate, but topic recognition accuracy deteriorates when keywords are absent or ambiguous terms are used
Solution Approach 1:
The system performs preliminary actions by building a knowledge base of topic descriptors and concepts before topic recognition occurs. This pre-processing enables the system to identify topics even when keywords are absent or ambiguous, as the knowledge base contains pre-defined relationships between concepts and topics that can be matched against conversation content.
Solution Approach 2:
The knowledge base acts as an intermediary between the conversation text and topic identification. Instead of directly matching keywords to topics, the system uses the knowledge base as a mediator that contains structured relationships between concepts, descriptors, and topics, enabling more accurate and nuanced topic recognition.
2Measurement precision
If a knowledge base with additional concepts is implemented, then topic recognition accuracy improves, but the complexity of maintaining and expanding the knowledge base increases
Solution Approach 1:
The knowledge base is designed to be universal and multi-functional, serving both as a repository of topic information and as a mechanism for automatic topic recognition. The structured format with descriptors and concepts allows the same knowledge base to handle multiple topics and conversation types, reducing the need for separate maintenance systems for different scenarios.
3Loss of information
If ambiguous terms are replaced with main topics through tagging, then information retrieval improves, but the processing time increases
Solution Approach 1:
The system applies partial action by selectively tagging only the most ambiguous or critical terms rather than attempting to replace every term. This approach improves information retrieval quality for key concepts while avoiding the excessive processing time that would result from comprehensive term replacement across the entire conversation.
Data Source
AI summary
A method, system and computer program product for topic recognition in a written conversation includes receiving, by one or more processors, a written conversation from a collaborative messaging system used by a user. The one or more processors perform a first search in the written conversation to find additional concepts associated with a topic stored in a knowledge base of the user. Each of the additional concepts includes a descriptor of the topic. In response to finding at least one additional concept associated with the topic in the written conversation, the one or more processors tag the at least one additional concept to the topic such that the at least one additional concept is replaced by the topic in the knowledge base.