Crisis Response Log Retrieval System
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Inexperienced users face challenges in accurately specifying risks and retrieving relevant past crisis response instances due to the complexity of inputting search conditions, and simply presenting full text responses does not enhance their experience or understanding.
Innovation Solution
A search device that stores past crisis response logs with problem, result, and response labels, allowing for structured retrieval of similar instances by matching input logs with corresponding task numbers, enabling users to associate similar portions and understand causal relationships.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If the user inputs a search condition to retrieve past crisis response instances, then the retrieval accuracy is improved, but the operation complexity increases for inexperienced users
Solution Approach 1:
The system automatically analyzes the input log and extracts search conditions without requiring manual input from the user. The search condition generation unit autonomously identifies risk factors and generates appropriate search queries, allowing the system to serve itself in the condition specification task.
Solution Approach 2:
The search condition generation unit acts as an intermediary between the user's input log and the search database. It translates the user's natural language input into structured search conditions, mediating the interaction and eliminating the need for users to directly construct complex search queries.
2Loss of information
If the full text of past response instances is presented to the user, then the information completeness is improved, but the information processing burden increases
Solution Approach 1:
The system extracts and presents only the relevant portions of past response instances that correspond to the risks identified in the input log. Instead of displaying complete full texts, it selectively extracts similar risk factors and their corresponding response actions, reducing information overload while maintaining completeness of relevant knowledge.
Solution Approach 2:
The past response instances are segmented into distinct risk factors and corresponding response actions. The system divides the comprehensive text into manageable segments based on similarity to the input log, allowing users to process information in discrete, relevant units rather than as a monolithic block.
3Adaptability or versatility
If structured labeling of past instances is implemented, then the know-how usability is improved, but the data processing complexity increases
Solution Approach 1:
The system performs preliminary analysis and labeling of past crisis response instances during the data preparation phase. Risk factors, response actions, and their relationships are pre-identified and structured before actual search operations, enabling rapid and adaptable querying without complex processing during runtime.
Data Source
AI summary
A search device has a know-how storing unit for storing a log indicating a past crisis response instance, one of a “problem” label indicating that the log is of a problem instance that is to be responded and solved, a “result” label indicating that the log is of a result brought about by a responding action, and a “response” label indicating that the log is of a responding action that has been performed to solve a problem, and a task number corresponding to the log, in association with each other. Further, there is a know-how search unit for repeating a search to retrieve a log similar to a first search key that is an input log from logs assigned the “problem” label in the know-how storing unit, and a search to retrieve a log similar to a second search key.