18226059. METHOD OF LEARNING NEURAL NETWORK, FEATURE SELECTION APPARATUS, FEATURE SELECTION METHOD, AND RECORDING MEDIUM simplified abstract (NEC Corporation)
METHOD OF LEARNING NEURAL NETWORK, FEATURE SELECTION APPARATUS, FEATURE SELECTION METHOD, AND RECORDING MEDIUM
Organization Name
Inventor(s)
Masanao Natsumeda of Tokyo (JP)
METHOD OF LEARNING NEURAL NETWORK, FEATURE SELECTION APPARATUS, FEATURE SELECTION METHOD, AND RECORDING MEDIUM - A simplified explanation of the abstract
This abstract first appeared for US patent application 18226059 titled 'METHOD OF LEARNING NEURAL NETWORK, FEATURE SELECTION APPARATUS, FEATURE SELECTION METHOD, AND RECORDING MEDIUM
Simplified Explanation
The abstract describes a method for learning a neural network that includes several layers: a feature selection layer, a feature extraction layer, a prediction layer, and a partial reconstruction layer. The method involves adjusting the weight parameter of the neural network based on the prediction accuracy and reconstruction error.
- The neural network includes a feature selection layer, feature extraction layer, prediction layer, and partial reconstruction layer.
- The feature selection layer selects a part of the input data.
- The feature extraction layer extracts a feature quantity from the selected input data.
- The prediction layer performs predictions based on the feature quantity.
- The partial reconstruction layer reconstructs the selected input data based on the feature quantity.
- The weight parameter of the neural network is adjusted based on the prediction accuracy and reconstruction error.
Potential Applications:
- Pattern recognition: The method can be used for pattern recognition tasks where selecting relevant features and accurately predicting patterns are important.
- Image processing: The method can be applied to image processing tasks such as object recognition or image classification.
- Natural language processing: The method can be used for tasks like sentiment analysis or text classification.
Problems Solved:
- Feature selection: The method addresses the problem of selecting relevant features from input data, improving the efficiency and accuracy of the neural network.
- Prediction accuracy: By adjusting the weight parameter based on prediction accuracy, the method aims to improve the overall accuracy of the neural network.
- Reconstruction error: The method considers the reconstruction error in the partial reconstruction layer, which helps in refining the selected input data.
Benefits:
- Improved efficiency: The feature selection layer helps in reducing the dimensionality of the input data, making the neural network more efficient.
- Enhanced accuracy: By adjusting the weight parameter based on prediction accuracy, the method aims to improve the overall accuracy of the neural network.
- Robustness: The partial reconstruction layer helps in reconstructing the selected input data, making the neural network more robust to noise or missing data.
Original Abstract Submitted
A method of learning a neural network, wherein the neural network includes: a feature selection layer for selecting a part of input data; a feature extraction layer for extracting a feature quantity on the basis of the selected input data; a prediction layer for performing a prediction on the basis of the feature quantity; and a partial reconstruction layer for reconstructing the selected input data on the basis of the feature quantity, and the method includes adjusting a weight parameter of the neural network on the basis of a prediction accuracy by the prediction layer and a reconstruction error in the partial reconstruction layer.
