Vehicle Neural Network Tainted-Result Handling for Data Errors
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing vehicle systems lack effective methods for handling errors in neural network processors, which can lead to safety risks and operational disruptions during real-world conditions.
Innovation Solution
A system comprising a neural network processor with an error detector and controller, capable of detecting data errors and signaling tainted results without terminating execution, allowing the system to continue operating while identifying and mitigating errors.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If the neural network processor terminates execution upon detecting a data error, then the reliability of the system is improved by preventing propagation of errors, but the productivity is reduced due to interruption of operation
Solution Approach 1:
The patent segments the error handling approach by distinguishing between different types of errors (data errors vs. other errors) and applying different responses. For data errors, the system identifies and isolates only the affected pending result rather than terminating the entire neural network execution, allowing other operations to continue while maintaining system reliability.
Solution Approach 2:
The system applies local quality by treating the erroneous data locally - when a data error is detected, only the specific pending result that depends on the erroneous data is identified and flagged as tainted. The rest of the neural network processor continues operating normally, thus maintaining high productivity while ensuring reliability through localized error containment.
2Productivity
If the system continues operation after detecting a data error, then the productivity is maintained, but the reliability deteriorates due to potential propagation of erroneous results
Solution Approach 1:
The patent implements feedback by continuously monitoring execution status and error conditions. When a data error is detected, the system provides feedback by identifying which pending results are tainted and preventing their use, thus maintaining reliability while allowing uninterrupted operation. The error detection unit and execution status indication unit work together to provide real-time feedback on data integrity.
3Reliability
If comprehensive error detection and handling mechanisms are implemented, then the reliability is improved, but the device complexity increases due to additional error detection and management components
Solution Approach 1:
The neural network processor implements self-service error handling through built-in error detection and status indication units that are integrated into the processor architecture. These units automatically detect data errors and identify affected results without requiring external intervention, thus improving reliability while minimizing the increase in device complexity through self-contained error management.
Data Source
AI summary
A system for handling errors in a neural network includes a neural network processor for executing a neural network associated with use of a vehicle. The neural network processor includes an error detector configured to detect a data error associated with execution of the neural network and a neural network controller configured to receive a report of the data error from the error detector. In response to receiving the report, the neural network controller is further configured to signal that a pending result of the neural network is tainted, without terminating execution of the neural network.