Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Cost-Aware Neural Early Stopping for Local Constraint OSD Decoders

Talha Akyildiz, Hessam Mahdavifar

Abstract

Local constraint ordered statistics decoding (LC-OSD) provides strong soft decision performance for short block length linear codes, but its practical cost is dominated by the number of tested error patterns (TEPs). This paper proposes a neural early stopping (NES) protocol for LC-OSD with explicit cost control through one trade-off parameter balancing frame error risk and search effort. The proposed approach is trained with frame error rate (FER)-aligned supervision at predefined checkpoints, and learns if additional search is still likely to improve the current best candidate. Later, stopping is decided by comparing predicted continuation need with a cost measured in TEPs. Experimental results across multiple code families show that the proposed protocol significantly reduces average TEP count with only marginal FER degradation, using a single global model for the range of all operating signal-to-noise ratios (SNRs).

Cost-Aware Neural Early Stopping for Local Constraint OSD Decoders

Abstract

Local constraint ordered statistics decoding (LC-OSD) provides strong soft decision performance for short block length linear codes, but its practical cost is dominated by the number of tested error patterns (TEPs). This paper proposes a neural early stopping (NES) protocol for LC-OSD with explicit cost control through one trade-off parameter balancing frame error risk and search effort. The proposed approach is trained with frame error rate (FER)-aligned supervision at predefined checkpoints, and learns if additional search is still likely to improve the current best candidate. Later, stopping is decided by comparing predicted continuation need with a cost measured in TEPs. Experimental results across multiple code families show that the proposed protocol significantly reduces average TEP count with only marginal FER degradation, using a single global model for the range of all operating signal-to-noise ratios (SNRs).
Paper Structure (12 sections, 16 equations, 4 figures, 1 algorithm)

This paper contains 12 sections, 16 equations, 4 figures, 1 algorithm.

Table of Contents

  1. Introduction
  2. System Model and Preliminaries
  3. Channel model and soft information
  4. Classical OSD
  5. LC-OSD
  6. Neural Early Stopping for LC-OSD
  7. Problem Formulation
  8. Stopping Rule
  9. Feature Representation
  10. Training Objective
  11. Numerical Results
  12. Conclusion

Figures (4)

  • Figure 1: Architecture of proposed NES framework for LC-OSD. Top: Runtime pipeline from channel LLRs to decoded codeword. Bottom: MLP estimator with three feature groups and asymmetric training loss.
  • Figure 2: FER and average TEP versus $E_b/N_0$ for $\mathcal{C}_1[128,64]$ and $\mathcal{C}_1[32,16]$.
  • Figure 3: FER and average TEP versus $E_b/N_0$ for $\mathcal{C}_2[128,64]$ and $\mathcal{C}_2[32,16]$.
  • Figure 4: FER and average TEP versus $E_b/N_0$ for $\mathcal{C}_3[128,64]$ and $\mathcal{C}_3[32,16]$.