Debug Log Vectorization for Malfunction Prediction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional methods for automatically identifying defects in machine tools through debug logs are inefficient due to reliance on similarity-based vectorization, which fails to accurately predict malfunctions and requires extensive manual processing.
Innovation Solution
An information processing apparatus that generates distributed representation vectors from debug logs, allowing for similarity calculations and malfunction analysis, incorporating preprocessing, integration determination, and external cooperation to improve identification accuracy and efficiency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Extent of automation
If similarity-based vectorization is used for debug log analysis, then the system can perform automatic identification, but the accuracy of malfunction prediction is insufficient and extensive manual processing is required
Solution Approach 1:
The patent transforms the parameter representation method from simple similarity-based vectorization to distributed representation vectors (embeddings) that capture semantic relationships between log entries. This parameter change in the mathematical representation enables more accurate malfunction prediction while maintaining automation, directly resolving the contradiction between automation extent and measurement precision.
2Measurement precision
If distributed representation vectors are generated for each line of debug logs, then identification accuracy is improved, but computational complexity and processing time increase
Solution Approach 1:
The patent segments the debug log analysis into line-level distributed representation generation followed by aggregation to device-level representations. This segmentation allows accurate line-level processing while managing computational complexity through hierarchical processing, resolving the contradiction between identification accuracy and device complexity.
Solution Approach 2:
The patent transitions from traditional vector space to high-dimensional distributed representation space, adding dimensional depth to capture semantic relationships. This dimensionality change enables more nuanced identification accuracy while the hierarchical aggregation structure manages the associated computational complexity.
3Measurement precision
If distributed representation vectors are used for debug log analysis, then malfunction analysis accuracy is enhanced, but processing time and computational resources increase
Solution Approach 1:
The patent performs preliminary generation of distributed representation vectors for individual log lines during normal operation, storing these representations for later aggregation and analysis. This preliminary action enables faster real-time malfunction analysis by pre-computing the computationally intensive vector representations, resolving the contradiction between analysis accuracy and processing time.
Data Source
AI summary
An information processing apparatus, a method, and a non-transitory recording medium. The information processing apparatus includes circuitry to acquire an operation log of a target device, generate a first vector of a distributed representation indicating the operation log, and calculate the first vector and a vector of a distributed representation based on another operation log to identify the first vector.