Automated Medical Data Selection for Privacy-Compliant ML Training
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The transfer of medical image data for machine learning is hindered by strict privacy regulations, requiring lengthy approval processes, and existing solutions like CometCloudCare are complex to implement and may not always desire distributed machine learning.
Innovation Solution
A system and method for collecting medical data that obfuscates patient identities, allowing privacy-compliant training data to be selected and transmitted from clinical sites to development sites, using an input interface for medical image data and label data, along with privacy policy data to limit data selection and ensure compliance, enabling machine learnable model training.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If medical image data is transferred outside clinical settings for machine learning, then machine learning model training can be performed with access to real patient data, but privacy regulations and approval processes hinder this transfer
Solution Approach 1:
The patent introduces an intermediary system that acts as a mediator between clinical data sources and machine learning models. This intermediary automatically selects and prepares training data according to privacy policies, enabling data transfer without manual approval processes while maintaining privacy compliance through automated criteria-based selection.
Solution Approach 2:
The system performs preliminary actions by pre-configuring privacy policies and selection criteria before data transfer. The automated selection process prepares data in advance according to established privacy rules, eliminating the need for lengthy retrospective approvals and enabling seamless data transfer for machine learning training.
2Reliability
If privacy policies are incorporated into the machine learning algorithm itself, then distributed machine learning with privacy preservation is enabled, but the system becomes complex to implement
Solution Approach 1:
The patent extracts privacy policy enforcement from the machine learning algorithm itself and places it in a separate data selection system. This separation allows the machine learning component to remain simple while privacy compliance is handled independently through automated data preparation, reducing overall system complexity while maintaining privacy preservation.
3Reliability
If manual oversight is used to ensure privacy compliance in data collection, then patient identity protection is maintained, but the data collection process becomes time-consuming and less efficient
Solution Approach 1:
The system implements self-service by enabling automated privacy-compliant data selection without manual intervention. The automated selection process independently evaluates and selects data according to privacy policies, maintaining patient identity protection through systematic criteria while dramatically improving data collection efficiency and eliminating bottlenecks associated with manual oversight.
Data Source
AI summary
A system (100) and computer-implemented method are provided for data collection for distributed machine learning of a machine learnable model. A privacy policy data (050) is provided defining computer-readable criteria for limiting a selection of medical image data (030) to a subset of the medical image data to obfuscate an identity of the at least one patient. The medical image data is selected based on the computer-readable criteria to obtain privacy policy-compliant training data (060) for transmission to another entity. The system and method enable medical data collection at clinical sites without requiring manual oversight, and enables such selections to be made automatically, e.g., based on a request for medical image data which may be received from outside of the clinical site.