Recurrent Neural Network ECC Decoding for Low-Power Error Correction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing error correction coding techniques in memory devices and wireless baseband circuitry are inefficient in terms of resource usage and processing complexity, leading to increased costs and longer development times, particularly in applications requiring ultra-low power consumption and latency like IoT and tactile internet.
Innovation Solution
The use of multi-layer neural networks and recurrent neural networks for decoding encoded data, which leverage nonlinear mapping and distributed processing capabilities to reduce errors introduced by noise, such as bit flips in non-volatile memory devices.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If complex error correction techniques are used, then error correction capability is improved, but area and power consumption increase
Solution Approach 1:
The patent replaces traditional mechanical/error-correction decoding algorithms with a neural network system that processes error correction in parallel, eliminating sequential processing bottlenecks and reducing power consumption while maintaining correction capability
Solution Approach 2:
The neural network is divided into multiple layers (input layer, hidden layers, output layer) that process error correction independently and in parallel, allowing distributed computation that reduces overall power consumption while maintaining comprehensive error correction
2Reliability
If complex error correction techniques are used, then error correction capability is improved, but device complexity increases
Solution Approach 1:
Complex iterative decoding algorithms are replaced with a neural network that performs error correction through forward propagation and backpropagation, simplifying the control logic and reducing processing complexity while maintaining or improving correction capability
Solution Approach 2:
The neural network is designed as a universal error correction system that can handle multiple types of errors and coding schemes through a single unified architecture, reducing the need for multiple specialized circuits and simplifying overall device complexity
3Device complexity
If traditional decoding methods are used, then implementation is simpler, but processing speed decreases
Solution Approach 1:
The neural network processes error correction through multiple parallel layers simultaneously, enabling pipelined and parallel computation that dramatically increases processing speed compared to sequential traditional methods while maintaining implementation feasibility
Solution Approach 2:
The neural network weights and structures are pre-trained offline to optimize error correction performance, allowing the deployed system to perform rapid inference without complex real-time training, thus achieving high speed while keeping implementation simple
Data Source
AI summary
Examples described herein utilize multi-layer neural networks, such as multi-layer recurrent neural networks to decode encoded data (e.g., data encoded using one or more encoding techniques). The neural networks and/or recurrent neural networks have nonlinear mapping and distributed processing capabilities which are advantageous in many systems employing the neural network decoders and/or recurrent neural networks. In this manner, neural networks or recurrent neural networks described herein are used to implement error correction coding (ECC) decoders.