Dual Neural Network Controller for Autonomous Machine Self-Correction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional neural networks lack the ability to effectively recognize and correct incorrect output signals and are unable to adaptively reduce mistakes with new information, leading to ineffective handling of incorrect outputs.
Innovation Solution
A controller system that includes a first neural network trained with initial data and a manager to re-train it incrementally using new data from sensors, along with a second neural network for real-time detection and correction, ensuring outputs remain within predetermined boundary conditions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If a conventional neural network is trained with initial data, then it can generate output signals, but it cannot recognize or correct incorrect outputs and cannot adaptively reduce mistakes with new information
Solution Approach 1:
The patent implements a feedback mechanism where the second neural network monitors the output of the first neural network and provides correction signals. The manager component uses feedback from detected incorrect outputs to re-train the first neural network incrementally with new data, enabling the system to adaptively reduce mistakes while maintaining reliable operation.
Solution Approach 2:
The patent introduces a second neural network as an intermediary monitoring system that independently verifies the output of the first neural network. This intermediary network acts as a watchdog that detects incorrect outputs and triggers corrective actions, resolving the contradiction by adding a verification layer without fundamentally changing the original network's function.
2Reliability
If a second neural network is added for real-time detection and correction, then reliability improves, but device complexity increases
Solution Approach 1:
The patent merges the monitoring and correction functions into a unified manager component that coordinates both neural networks. The manager integrates the training data collection, incremental re-training, and output verification processes into a single control unit, reducing overall system complexity while maintaining the dual-network architecture for reliability.
Solution Approach 2:
The patent implements preliminary action by pre-training the second neural network to recognize patterns of incorrect outputs before actual operation. The manager also pre-establishes the incremental re-training mechanism with buffered training data, so that when incorrect outputs are detected, the correction process can begin immediately without adding computational complexity during real-time operation.
Data Source
AI summary
An apparatus having components implemented on one or more solid-state chips. The apparatus includes an input device constructed to generate an input data value (input value), and a neural network implemented on solid-state chips trained to generate an output to control the apparatus by processing the input value. The apparatus also includes another neural network implemented on solid-state chips and configured to receive the output from the neural network. The another neural network is trained to determine whether the output of the neural network corresponds to a predetermined condition and generate a control output from the output of the neural network. The apparatus includes a processor configured receive the control output from the aforementioned another neural network, and in response to the control output indicating the output of the first neural network corresponds to a predetermined condition, and control an operation of the neural network. Corresponding methods are also disclosed.