Sensor Data Combination Optimization for Machine Learning
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current machine learning systems lack the ability to systematically evaluate the contribution of various sensor data and optimize sensor configurations, leading to inefficiencies in performance, cost, and processing time.
Innovation Solution
An information processing apparatus that includes a machine learning system capable of evaluating the contribution of sensor data by generating combinations, setting expected performance criteria, and optimizing sensor configurations through a performance evaluation process, which involves selecting relevant sensor data and adjusting neural network settings to achieve desired performance metrics.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If multiple sensor data are inputted to machine learning apparatus, then learning accuracy and recognition performance are improved, but device cost and processing time increase
Solution Approach 1:
The patent extracts and identifies the essential sensor data that contribute most to learning accuracy from multiple sensor inputs. By systematically evaluating the contribution rate of each sensor data type, the system extracts only the necessary sensors needed to achieve required performance, removing redundant sensors to reduce device cost and complexity while maintaining learning accuracy.
2Measurement precision
If multiple sensor data are inputted to machine learning apparatus, then learning accuracy and recognition performance are improved, but processing time and computational load increase
Solution Approach 1:
The patent extracts only the essential sensor data that contribute most to recognition performance. By evaluating the contribution rate of each sensor data type and selecting only those with high contribution rates, the system reduces the volume of data to be processed while maintaining recognition performance, thereby reducing processing time and computational load.
3Device complexity
If sensor data with low contribution rate are identified and removed, then device cost and processing time are reduced, but learning accuracy may deteriorate
Solution Approach 1:
The patent implements a feedback mechanism where the machine learning apparatus evaluates the contribution rate of each sensor data type to the learning results. This evaluation feedback is used to systematically identify which sensors can be removed without significantly affecting learning accuracy, enabling data-driven optimization of sensor configuration that balances cost reduction with performance maintenance.
Solution Approach 2:
The patent applies partial action by selectively removing only those sensor data types whose contribution rate falls below a predetermined threshold. This partial removal approach ensures that learning accuracy is maintained at an acceptable level while still achieving cost and complexity reduction, rather than removing all non-essential sensors aggressively.
Data Source
AI summary
An information processing apparatus generates multiple combinations of sensor data inputted to a machine learning apparatus, inputs the combinations of sensor data to the machine learning apparatus, and generates a recognizer corresponding to each of the combinations of sensor data. Further, the performance of the recognizers is evaluated in accordance with expected performance required for the recognizers, and the combinations of sensor data corresponding to the recognizers satisfying the expected performance are outputted. Thus, the rates of contribution of two or more pieces of sensor data inputted to the machine learning apparatus are evaluated, and the configuration of sensors is optimized.