Equal-Unit Neural Network Architecture for Mobile Memory Reduction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Artificial neural networks (ANNs) are computationally complex and require significant storage, making it impractical to run them on mobile devices for predictive language models, which limits their use in mobile electronic devices due to resource constraints.
Innovation Solution
A mobile-friendly ANN architecture is implemented with an input layer, a first hidden layer, a second hidden layer of equal size to the input layer, and an output layer, using the same matrix for encoding and decoding, reducing storage and memory requirements while maintaining accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a traditional artificial neural network is used for predictive language modeling, then prediction accuracy is improved, but storage requirements and computational complexity increase significantly
Solution Approach 1:
The neural network is segmented into multiple layers (input layer, hidden layers, output layer) with specific unit configurations. The second hidden layer is designed to have the same number of units as the input layer, creating a structured segmentation that reduces redundant computations and storage requirements while maintaining predictive accuracy.
Solution Approach 2:
The network architecture parameters are optimized by setting the number of units in the second hidden layer equal to the input layer units. This parameter change reduces the overall complexity and storage requirements of the network while preserving its ability to perform accurate language prediction.
2Adaptability or versatility
If a traditional artificial neural network is deployed on mobile devices, then predictive language modeling capability is improved, but device memory consumption increases
Solution Approach 1:
The network is divided into functional layers with the second hidden layer matching the input layer size, creating an efficient structure that reduces memory footprint while maintaining full predictive language modeling functionality on mobile devices.
Solution Approach 2:
The optimized network architecture serves multiple functions: it performs accurate language prediction, consumes less memory, and can be deployed on resource-constrained mobile devices, making the solution universally applicable across different device types.
3Quantity of substance
If neural network complexity is reduced for mobile deployment, then storage requirements decrease, but prediction accuracy may be compromised
Solution Approach 1:
By changing the architectural parameter to make the second hidden layer unit count equal to the input layer unit count, the network achieves a compact structure that reduces storage requirements while preserving prediction accuracy through optimized information flow and feature representation.
Data Source
AI summary
The present invention relates to an improved artificial neural network for predicting one or more next items in a sequence of items based on an input sequence item. The improved artificial neural network has greatly reduced memory requirements, making it suitable for use on electronic devices such as mobile phones and tablets. The invention includes an electronic device on which the improved artificial neural network operates, and methods of predicting the one or more next items in the sequence using the improved artificial neural network.