Medical Information Processing Apparatus Model Update Strategy

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Current medical decision support technologies, such as supervised learning, require large amounts of data and cannot demonstrate causal sequences, leading to inefficiencies and limitations in clinical decision-making, particularly in personalized medicine, and are hindered by the costs and time-consuming nature of randomized controlled trials.

Innovation Solution

A medical information processing apparatus and method that updates models for calculating effect evaluation values by separating parameter and structure updates, allowing for more frequent parameter updates while retaining the model structure, and adjusting the structure less frequently based on accumulated observation data, enabling adaptive and efficient medical decision refinement.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Adaptability or versatility

If the model structure is updated frequently to improve adaptability and refinement capability, then the degree of realizable refinement of medical decisions is improved, but the computational complexity and time consumption increase

Engineering Contradiction:
Improverefinement capabilityVSAvoidmodel update complexity
Core Design Contradiction:
Adaptability or versatilityVSDevice complexity

Solution Approach 1:

The patent segments the model into two distinct components: parameters and structure. Parameters are updated frequently based on new observation data to capture current patterns, while the structure is updated less frequently to maintain stability. This segmentation allows independent optimization of each component's update frequency, resolving the contradiction between adaptability and complexity.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent implements dynamic update strategies where the update frequency of parameters and structure differs. Parameters undergo frequent updates to adapt to changing medical patterns, while structure updates are spaced out to avoid excessive complexity. This dynamic approach allows the system to balance adaptability with computational feasibility.

Inventive Principle:
Principle #15Dynamics

2Ease of operation

If the model structure is simplified to reduce computational complexity, then the ease of operation is improved, but the inference accuracy deteriorates

Engineering Contradiction:
Improvecomputational efficiencyVSAvoidinference accuracy
Core Design Contradiction:
Ease of operationVSMeasurement precision

Solution Approach 1:

By segmenting the model into parameters and structure, the patent allows the structure to remain relatively simple and stable while parameters capture the necessary complexity for accurate inference. This segmentation enables the system to maintain computational efficiency through simple structure while achieving high accuracy through parameter optimization.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent relies on parameter changes rather than structural changes to achieve model adaptation. By updating parameters frequently based on observation data, the system maintains high inference accuracy without the computational burden of frequent structural modifications, thus preserving ease of operation.

Inventive Principle:
Principle #35Parameter changes

3Measurement precision

If parameter updates are performed frequently to improve inference accuracy, then the measurement precision is improved, but the device complexity increases

Engineering Contradiction:
Improveinference accuracyVSAvoidupdate frequency management
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The patent segments update operations into two distinct processes: parameter updates (frequent) and structure updates (infrequent). This segmentation allows the system to manage update frequency by treating parameters and structure differently, reducing the complexity of update management while maintaining high inference accuracy through frequent parameter optimization.

Inventive Principle:
Principle #1Segmentation

4Adaptability or versatility

If the model structure is updated frequently to improve adaptability, then the refinement capability is improved, but the loss of time increases

Engineering Contradiction:
Improverefinement capabilityVSAvoidupdate time
Core Design Contradiction:
Adaptability or versatilityVSLoss of time

Solution Approach 1:

The patent segments update operations into parameter updates (frequent, fast) and structure updates (infrequent, more time-consuming). This segmentation allows the system to achieve adaptability through frequent parameter updates without the time penalty of frequent structural modifications, thus reducing overall time loss while maintaining refinement capability.

Inventive Principle:
Principle #1Segmentation

Data Source

PatentEP4160619A1Medical information processing apparatus
Publication Date: 2023.04.05 CANON MEDICAL SYST CORP
  • EP4160619A1 patent drawingFigure 1
  • EP4160619A1 patent drawingFigure 2
  • EP4160619A1 patent drawingFigure 3

AI summary

According to one embodiment, a medical information processing apparatus includes processing circuitry which updates a model for calculating an effect evaluation value for a medical decision. The processing circuitry updates a parameter of the model while retaining the structure of the model so that the structure of the model is updated less frequently than the parameter.