Chunk-Based AI Simulation for Secure Anomaly Detection Tuning
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The challenge in condition-based maintenance of mechanical systems is the difficulty in sharing user-specific data for AI training due to concerns over secret information disclosure and the need for domain knowledge, leading to inappropriate AI algorithm selection and inefficient data analysis.
Innovation Solution
A simulation apparatus and method that allows users to verify AI effectiveness using unsupervised training with their own data, employing a machine learning model with preprocessing units and a three-layer neural network for anomaly detection, enabling in-place analysis and reducing the risk of data exposure.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If user-specific data is shared for AI training, then AI algorithm effectiveness is improved, but risk of secret information disclosure increases
Solution Approach 1:
The patent creates a simulation environment that copies the essential characteristics of real AI training systems. Users can train machine learning models using simulated data that replicates the behavior and patterns of actual operational data without exposing sensitive information. The simulation apparatus generates synthetic training datasets that preserve the statistical properties and relationships needed for effective model training while eliminating security risks associated with sharing real data.
2Measurement precision
If domain knowledge is required for AI training, then AI algorithm selection accuracy is improved, but ease of operation deteriorates
Solution Approach 1:
The simulation apparatus enables users to independently verify AI algorithm effectiveness using their own operational data in a controlled simulation environment. Users can perform self-assessment of different machine learning models without requiring external domain experts. The system provides tools for users to configure training parameters, execute experiments, and analyze results autonomously, thereby reducing dependency on specialized knowledge while maintaining accurate algorithm selection.
3Power
If AI training is performed externally, then computational resources are improved, but loss of time in data preparation and verification increases
Solution Approach 1:
The simulation apparatus allows users to perform preliminary verification of AI algorithm effectiveness before committing to external training processes. Users can conduct rapid experiments with simulated data to assess model performance, validate data quality, and optimize training configurations in advance. This preliminary action eliminates time-wasting iterations later when working with actual external training systems, as users have already identified promising approaches and potential issues in the simulation environment.
Data Source
AI summary
In the simulation apparatus, a model setting unit executes setting related to a chunk as a batch block of data in a case where data are sequentially entered into a machine learning model on a basis of a loaded sequence of data. A model computing unit executes computations of unsupervised training and computations of prediction by sequentially entering the chunk into the machine learning model. The model setting unit sets a range of data for use in computations of unsupervised training from within the sequence of data by a chunk designated with input by an operation input portion.