Medical Information Processing Apparatus Causal Inference Model
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In individualized medical care, it is challenging to estimate therapeutic effects due to the scarcity of training samples for machine learning, and existing methods lack integration of both machine learning and past medical knowledge to improve accuracy.
Innovation Solution
A medical information processing apparatus that acquires training samples with feature amounts, type labels, and effect labels, and a knowledge base independent of these samples, assigns knowledge labels, and trains a causal inference model to infer the effects of medical events, incorporating both machine learning and medical knowledge.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If machine learning based on training samples is used to estimate therapeutic effects, then the model can learn from data, but the accuracy is insufficient when training samples are scarce
Solution Approach 1:
The patent combines machine learning models with knowledge graphs that encode medical knowledge. The system integrates both data-driven learning from training samples and knowledge-driven reasoning from the knowledge graph, allowing accurate therapeutic effect estimation even when training samples are limited. The knowledge graph provides additional structural information and medical expertise that compensates for insufficient training data.
Solution Approach 2:
The knowledge graph serves as an intermediary between the limited training samples and the therapeutic effect estimation task. It mediates by providing additional contextual information, medical relationships, and domain knowledge that bridge the gap between scarce data and accurate predictions, enabling the system to achieve high accuracy without requiring large numbers of training samples.
2Productivity
If only training samples are used for machine learning, then the model can be trained efficiently, but existing medical knowledge is not utilized
Solution Approach 1:
The system merges traditional machine learning approaches that prioritize training efficiency with knowledge graphs that capture medical expertise. By integrating these two paradigms, the system maintains efficient training processes while simultaneously incorporating rich medical knowledge from the knowledge graph, thus avoiding the loss of valuable domain information.
Solution Approach 2:
The knowledge graph is constructed and populated with medical knowledge in advance, before the training process begins. This preliminary action allows the system to have ready-accessible medical expertise during training and inference, eliminating the need to relearn basic medical relationships from scratch and preserving valuable domain knowledge throughout the training process.
Data Source
AI summary
A medical information processing apparatus acquires multiple training samples. Each of the training samples includes a feature amount representing a condition of a subject, a type label of an event performed on the subject, and an effect label of the event. The apparatus acquires a knowledge base independent from the training samples. The processing circuitry assigns a knowledge label to at least one training sample among the training samples based on the knowledge base. The apparatus trains, based at least on the at least one training sample to which the knowledge label is assigned, a model that infers an effect of each type of an event. The at least one training sample to which the knowledge label is assigned includes the feature amount, the type label, the effect label, and the knowledge label.