Question Generating Apparatus for Intelligent Diagnosis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current intelligent inquiring diagnosis systems require multiple interactions with patients to collect comprehensive symptoms and signs information, leading to delayed diagnosis and inefficient use of medical resources.
Innovation Solution
A question generating apparatus and method that calculates the information value of candidate questions based on their correlation with the dialog context, using recurrent neural networks to select questions that maximize information gain, thereby reducing the number of interactions needed for accurate diagnosis.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If multiple interactions with patients are conducted to collect comprehensive symptoms and signs information, then diagnostic accuracy is improved, but diagnosis time and resource consumption increase
Solution Approach 1:
The patent replaces the mechanical interaction process (multiple sequential questions) with an information theory-based selection mechanism. By calculating information values using probability models and selecting questions that maximize information gain, the system achieves comprehensive diagnostic information collection in fewer interactions, thus reducing diagnosis time while maintaining accuracy
Solution Approach 2:
The patent changes the parameter selection criterion from traditional rule-based or fixed question sequences to dynamic information value calculation. By continuously evaluating the information value of candidate questions based on current dialog context and selecting questions with maximum expected information gain, the system optimizes the questioning strategy to achieve diagnostic goals faster
2Reliability
If multiple interactions with patients are conducted to collect comprehensive symptoms and signs information, then diagnostic accuracy is improved, but resource efficiency deteriorates
Solution Approach 1:
The patent replaces inefficient mechanical interaction processes with an information-theoretic optimization system. By calculating information values and selecting questions that maximize information gain per interaction, the system achieves comprehensive diagnostic information collection in fewer interactions, thus improving resource efficiency while maintaining diagnostic accuracy
Solution Approach 2:
The system performs self-optimization of the questioning strategy by automatically calculating information values and selecting the most valuable questions without human intervention. This autonomous optimization ensures that each interaction maximizes information gain, improving overall resource efficiency of the diagnostic process
3Productivity
If questions are selected based on information value calculation, then information collection efficiency is improved, but system complexity increases
Solution Approach 1:
The patent introduces information value as an intermediary metric that bridges the gap between dialog context and question selection. By calculating the information value of candidate questions based on their expected information gain and selecting questions with maximum values, the system achieves efficient information collection while managing complexity through a clear mathematical framework
Data Source
AI summary
The present disclosure relates to a question generating method and apparatus, an inquiring diagnosis system, and a computer readable storage medium. The question generating apparatus includes at least one processor, the at least one processor being configured to: acquire a candidate question set Q; calculate an information value of each candidate question in the candidate question set Q; and generate at least one question according to the information value of each candidate question.