Neural-Network-Optimized Degree-Specific Weights for LDPC MinSum Decoding

Abstract

Neural Normalized MinSum (N-NMS) decoding delivers better frame error rate (FER) performance on linear block codes than conventional normalized MinSum (NMS) by assigning dynamic multiplicative weights to each check-to-variable message in each iteration. Previous N-NMS efforts have primarily investigated short-length block codes (N < 1000), because the number of N-NMS parameters to be trained is proportional to the number of edges in the parity check matrix and the number of iterations, which imposes am impractical memory requirement when Pytorch or Tensorflow is used for training. This paper provides efficient approaches to training parameters of N-NMS that support N-NMS for longer block lengths. Specifically, this paper introduces a family of neural 2-dimensional normalized (N-2D-NMS) decoders with with various reduced parameter sets and shows how performance varies with the parameter set selected. The N-2D-NMS decoders share weights with respect to check node and/or variable node degree. Simulation results justify this approach, showing that the trained weights of N-NMS have a strong correlation to the check node degree, variable node degree, and iteration number. Further simulation results on the (3096,1032) protograph-based raptor-like (PBRL) code show that N-2D-NMS decoder can achieve the same FER as N-NMS with significantly fewer parameters required. The N-2D-NMS decoder for a (16200,7200) DVBS-2 standard LDPC code shows a lower error floor than belief propagation. Finally, a hybrid decoding structure combining a feedforward structure with a recurrent structure is proposed in this paper. The hybrid structure shows similar decoding performance to full feedforward structure, but requires significantly fewer parameters.

Joint-Way Compression for LDPC Neural Decoding Algorithm With Tensor-Ring Decomposition

“…The weight coefficient sharing algorithm, similar to the pruning and compression algorithm, is a low-complexity technique commonly used in channel neural decoding [5], [9], [10]. We apply two typical weight coefficient sharing methods [9], namely temporal weight sharing and spatial weight sharing, to the LDPC decoding algorithm based on model-driven deep learning and proposed a low complexity LDPC neural decoding algorithm.…”

“…The authors in [9] proposed a weight sharing method for learning BP decoding algorithm, the weight sharing method can be divided into two types, namely temporal weight sharing and spatial weight sharing. The authors in [10] applied weight sharing to the decoding algorithm of long LDPC codes, optimized the parameters according to the degree of nodes, and proposed Neural 2-dimensional Normalized decoders. The authors in [11] used the weight sharing algorithm in the proposed NNMS decoding algorithm, and then quantized the parameters of the neural decoding network using the bit quantization algorithm.…”

Research on Data Link Channel Decoding Optimization Scheme for Drone Power Inspection Scenarios

“…To address the issue of high decoding complexity caused by multiple weight parameters, eliminating redundant parameters is crucial. Wang et al [32] mentioned a neural two-dimensional normalized minimum sum decoder with a simplified parameter set, which allows for assigning the same weight to a class of similar messages. Lian et al [33] introduced a parameter sharing scheme in the weighted belief propagation [27] decoder, where certain edges share the same weights to reduce storage and computational burden.…”

Boost decoding performance of finite geometry LDPC codes with deep learning tactics

“…Furthermore, for a class of protograph-based 5G LDPC codes, whose structure is quasi cyclic, [21], [22] explored to apply machine learning techniques with fully sharing edge weights for long block codes. Besides that, for a class of irregular LDPC codes, it was proposed in [23] to assign and share weights in terms of its degree distribution for the NNMS decoder.…”