CRISP: Curriculum based Sequential Neural Decoders for Polar Code Family

TL;DR

CRISP is a novel curriculum-based sequential neural decoder for polar and PAC codes that leverages information-theoretic insights to achieve near-optimal reliability, outperforming traditional decoders especially in short blocklength regimes.

Contribution

Introduces CRISP, the first data-driven neural decoder for PAC codes, with a principled curriculum guiding training to improve reliability over existing methods.

Findings

CRISP outperforms successive-cancellation decoders on Polar(32,16) and Polar(64,22).

CRISP achieves near-optimal performance on PAC(32,16) codes.

The curriculum design is critical for the decoder's success.

Abstract

Polar codes are widely used state-of-the-art codes for reliable communication that have recently been included in the 5th generation wireless standards (5G). However, there remains room for the design of polar decoders that are both efficient and reliable in the short blocklength regime. Motivated by recent successes of data-driven channel decoders, we introduce a novel $\textbf{C}$ur$\textbf{RI}$culum based $\textbf{S}$equential neural decoder for $\textbf{P}$olar codes (CRISP). We design a principled curriculum, guided by information-theoretic insights, to train CRISP and show that it outperforms the successive-cancellation (SC) decoder and attains near-optimal reliability performance on the Polar(32,16) and Polar(64,22) codes. The choice of the proposed curriculum is critical in achieving the accuracy gains of CRISP, as we show by comparing against other curricula. More notably, CRISP can be readily extended to Polarization-Adjusted-Convolutional (PAC) codes, where existing SC decoders are significantly less reliable. To the best of our knowledge, CRISP constructs the first data-driven decoder for PAC codes and attains near-optimal performance on the PAC(32,16) code.