Machine Learning Audit Trail for Time Series Tampering Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing systems lack the ability to effectively audit the results of machine learning models, leading to issues such as data tampering and corruption, which can result in regulatory fines and adversarial relationships between industries and governments.
Innovation Solution
A method that accesses original time series data and machine learning estimates, generates reconstituted data by reversing the machine learning model operations, and compares these to detect tampering or corruption using a sequential probability ratio test, while creating a tamper-proof data set that can be audited.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If machine learning models are used to process sensor data, then productivity and automation are improved, but the ability to audit results and detect data tampering deteriorates
Solution Approach 1:
The system performs preliminary actions by creating tamper-proof copies of sensor data and storing audit trails before the data is processed by machine learning models. This allows subsequent auditing without requiring hardware upgrades or disrupting the automated processing pipeline.
Solution Approach 2:
The invention creates copies of the original sensor data in a tamper-proof format and stores them separately from the processed data. These copies serve as reference for auditing purposes, allowing verification of data integrity without affecting the machine learning processing workflow.
2Reliability
If data tampering detection is implemented, then reliability and auditability are improved, but device complexity increases
Solution Approach 1:
The system introduces an intermediary audit trail that records transformations and processing steps between original sensor data and machine learning outputs. This intermediary layer enables tampering detection without requiring complex changes to the existing machine learning infrastructure.
Solution Approach 2:
The auditing function is segmented into separate components: data capture, tamper-proof storage, audit trail generation, and comparison analysis. This segmentation allows the auditing mechanism to be added independently without increasing the complexity of the core machine learning system.
3Measurement precision
If comprehensive auditing is performed, then measurement precision and tampering detection are improved, but loss of time increases
Solution Approach 1:
Audit trails and tamper-proof data copies are prepared in advance during data collection, so that when auditing is needed, the system can perform rapid comparison without time-consuming data reconstruction or analysis.
Solution Approach 2:
The system replaces complex mechanical or computational auditing processes with straightforward digital comparison of hash values and audit trail records, dramatically reducing the time required to verify data integrity while maintaining high detection precision.
Data Source
AI summary
Systems, methods, and other embodiments associated with auditing the results of a machine learning model are described. In one embodiment, a method accesses original time series data and machine learning estimates of the original time series data. The method generates reconstituted time series data from the machine learning estimates by reversing operations of a machine learning model trained for generating the machine learning estimates from the original time series data. The method detects tampering (or corruption) in the original time series data based on a difference between the original time series data and reconstituted time series data. And, the method generates an electronic verification report that indicates whether the tampering (or corruption) is detected in the original time series data.