Meeting Support System with Dynamic Topic Control
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing meeting support technologies fail to effectively prevent meetings from going off-topic from a predetermined subject, especially when all speakers approve off-topic content, and may hinder progress in brainstorming sessions or fail to consider actual spoken content.
Innovation Solution
A meeting support system that determines a first feature value based on pre-registered meeting information, extracts and evaluates words or phrases from speech using natural language processing, and compares them to an allowable range specified by weights corresponding to the meeting type and elapsed time to determine if the discussion is on-topic.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If strict topic control is applied to prevent off-topic discussion, then meeting subject compliance is improved, but meeting progress is hindered in brainstorming sessions
Solution Approach 1:
The system dynamically adjusts the allowable off-topic range based on meeting type. For brainstorming meetings, a larger allowable range is set to permit creative off-topic discussions, while for decision-making meetings, a stricter range is applied. This dynamic adjustment resolves the contradiction by making the topic control flexible rather than fixed, allowing both strict control and progress to be achieved in appropriate contexts.
Solution Approach 2:
The system changes the parameter of allowable off-topic range according to meeting type and elapsed time. By modifying this parameter dynamically, the system can maintain high topic control precision when needed while avoiding hindrance to meeting progress in creative sessions. The parameter change approach allows the same system to serve multiple meeting objectives effectively.
2Ease of manufacture
If off-topic detection is based on predetermined keywords, then detection simplicity is improved, but accuracy in determining actual off-topic speech is reduced
Solution Approach 1:
The system replaces simple keyword matching (mechanical approach) with semantic analysis using natural language processing. This substitution enables accurate understanding of whether speech is truly off-topic by analyzing the meaning and context, rather than merely checking for keyword presence. The NLP-based approach maintains detection simplicity while dramatically improving accuracy.
3Device complexity
If uniform topic control is applied to all meeting types, then system simplicity is improved, but adaptability to different meeting contexts is reduced
Solution Approach 1:
The system applies different allowable off-topic ranges to different meeting types (local quality adjustment). Instead of using a uniform control mechanism for all meetings, the system tailors the topic control parameters to each meeting's specific needs. This allows the system to remain relatively simple while achieving high adaptability across diverse meeting contexts.
Data Source
AI summary
There is provided a meeting support system that supports the progress of a meeting, the meeting support system including a hardware processor that: determines a first feature value based on information about a meeting, the first feature value corresponding to a range of a subject of a meeting, the information about a meeting having been registered in advance; stores a weight, the weight corresponding to each type of meeting; extracts a word or phrase from speech of participants in a meeting; determines a second feature value corresponding to the extracted word or phrase; and determines whether a speech is off topic from a subject of a meeting, depending on whether the second feature value is included in an allowable range, the allowable range being specified by the first feature value and the weight corresponding to a type of an ongoing meeting.