Neural Product Autoencoder for Large-Dimension Error Correction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing technologies face challenges in training neural networks for channel encoding and decoding, particularly for large code dimensions, due to dimensionality issues and the need for huge networks with excessively large numbers of learnable parameters.
Innovation Solution
The implementation of a neural product coding system using multiple neural encoder and decoder stages, where each stage is configured as a separate neural network, allows for the construction of large neural codes from smaller code components. This system employs a product autoencoder architecture for joint training of encoder and decoder neural networks.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Quantity of substance
If a single large neural network is used for channel encoding and decoding, then the code dimension k can be large, but the number of learnable parameters becomes excessively large
Solution Approach 1:
The patent divides a large neural network into multiple smaller neural networks arranged in a product code structure. Each smaller network handles a portion of the encoding/decoding task, and their combined output achieves the desired large code dimension. This segmentation reduces the number of learnable parameters while maintaining the ability to encode large code dimensions through the product structure of the smaller networks.
2Ease of manufacture
If classical error correction codes are used, then implementation is simple, but bit error rate performance is inferior to neural network approaches
Solution Approach 1:
The patent applies different characteristics to different parts of the system: smaller neural networks are used in specific positions within the product code structure, allowing each component to be trained independently with manageable complexity while achieving superior overall bit error rate performance compared to uniform classical codes.
Data Source
AI summary
A processing circuit implements: an encoder configured to: supply k symbols of original data to a neural product encoder including M neural encoder stages, a j-th neural encoder stage including a j-th neural network configured by j-th parameters to implement an (nj,kj) error correction code (ECC), where nj is a factor of n and kj is a factor of k; and output n symbols representing the k symbols of original data encoded by an error correcting code; or a decoder configured to supply n symbols of a received message to a neural product decoder including neural decoder stages grouped into a l pipeline stages, an i-th pipeline stage of the neural product decoder including M neural decoder stages, a j-th neural decoder stage comprising a j-th neural network configured by j-th parameters to implement an (nj,kj) ECC; and output k symbols decoded from the n symbols of the received message.