Immutable Audit Trail for Machine Learning Models
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Machine learning algorithms in banking lack transparency and accountability in adhering to Fair Lending rules, as they focus solely on data patterns without ensuring compliance with public policies, potentially leading to biased decisions.
Innovation Solution
An apparatus and method utilizing an immutable storage facility, such as blockchain or write-once-read-many storage, to archive machine learning models and customer data, ensuring that the reasoning behind decisions is stored in an unmodifiable and auditable format, providing a defensible trail for compliance verification.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If machine learning algorithms are used to detect fraud and make lending decisions, then fraud detection capability and decision-making efficiency are improved, but transparency and accountability regarding Fair Lending compliance deteriorate
Solution Approach 1:
The patent applies preliminary action by storing the reasoning and decision-making process in an immutable archive before the final decision is made. This allows the system to maintain fraud detection capabilities while preserving transparency, as the archived reasoning can be reviewed later to verify Fair Lending compliance without interfering with the real-time decision-making efficiency.
2Measurement precision
If machine learning models continuously update based on new data, then fraud detection accuracy is improved, but the ability to audit and verify compliance with Fair Lending rules deteriorates
Solution Approach 1:
The patent applies copying by creating immutable copies of the machine learning model's reasoning and decision-making process at the time of each decision. These copies are stored in an immutable archive, allowing the model to continuously update and improve fraud detection accuracy while maintaining reliable audit trails that capture the state of the model at specific points in time for compliance verification.
3Reliability
If immutable storage is used to archive machine learning models and decisions, then accountability and compliance verification are improved, but storage complexity and infrastructure requirements worsen
Solution Approach 1:
The patent applies the intermediary principle by introducing an immutable archive as a mediator between the machine learning model and the audit/compliance verification process. This archive serves as an intermediate layer that stores decision-making reasoning in a tamper-proof manner, improving accountability without requiring complex changes to the core machine learning infrastructure. The archive acts as a separate, specialized storage component that simplifies the overall system architecture.
Data Source
AI summary
An apparatus and method for providing an immutable audit trail for machine learning applications is described herein. The audit trail is preserved by recording the machine learning models and data in a data structure in immutable storage such as a WORM device, a cloud storage facility, or in a blockchain. The immutable audit trail is important for providing bank auditors with the reasons for lending or account opening reasons, for example. A graphical user interface is described to allow the archive of machine learning models to be viewed.