RNN Encoder-Decoder for Clinical Trial Criteria Curation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current data mining methods for determining eligible patients for clinical trials are prone to inaccuracies due to incorrect, unclear, or logically flawed textual criteria, leading to mistrustworthy results and under-enrollment of eligible patients, as they require manual curation and are not machine-readable.
Innovation Solution
Employing a pair of trained recurrent neural networks (RNNs) as an encoder and decoder to preprocess unrefined textual criteria into curated, machine-readable output criteria, which are then ingested by a machine learning model using natural language processing to identify eligible candidates for clinical trials.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If manual data mining methods are used to process textual criteria, then human expertise can interpret ambiguous statements, but the process becomes impractical for large data sets and remains prone to inaccuracy
Solution Approach 1:
The system employs trained recurrent neural networks that automatically process and curate textual criteria without requiring manual human intervention. The encoder-decoder RNN architecture self-corrects errors and formats criteria into machine-readable forms, enabling the system to serve itself for large-scale data processing while maintaining accuracy
Solution Approach 2:
The patent replaces manual mechanical data mining processes with automated neural network-based processing. The encoder-decoder RNN system substitutes human experts' manual interpretation work with machine learning models that can process large volumes of textual criteria efficiently and consistently
2Productivity
If inaccurate textual criteria are ingested into the system, then more raw data can be processed, but the results become unreliable and the number of identified eligible patients decreases
Solution Approach 1:
The system performs preliminary cur curation of textual criteria before they are ingested into the patient identification process. The encoder-decoder RNN processes and corrects textual criteria in advance, converting them into accurate machine-readable formats, which ensures that subsequent patient identification operates on reliable data
Solution Approach 2:
The encoder-decoder RNN system acts as an intermediary layer between raw textual criteria and the patient identification algorithm. This intermediate processing step transforms unrefined textual statements into curated, accurate criteria, mediating between the raw data and the final analysis to ensure reliability
3Measurement precision
If manual curation is performed to correct defects in textual criteria, then accuracy can be improved, but the process becomes impractical for large data sets
Solution Approach 1:
The patent replaces manual curation operations with automated recurrent neural networks. The encoder-decoder architecture performs what would otherwise require human experts to review and correct each textual criterion, automatically handling large data sets without increasing operational complexity for users
Solution Approach 2:
The neural network system automatically curates its own input data by detecting and correcting errors in textual criteria. The encoder processes the input and the decoder generates corrected output, creating a self-service mechanism that maintains accuracy without requiring manual intervention for each data point
Data Source
AI summary
A method and apparatus for providing curated criteria to identify one or more candidates for a clinical trial is disclosed. A computer processor identifies a first input criterion for the clinical trial. The processor employs a trained first recurrent neural network (RNN) configured as an encoder to encode the first input criterion. The encoder extracts key features of the medical condition of the patient. The processor employs a trained second RNN configured as a decoder to generate a curated output criterion by processing the encoded first input criterion based on the derived key features. The processor employs a machine learning model to ingest the curated output criterion to identify the one or more candidates for the clinical trial.