We propose an unsupervised deep learning-based decoding scheme that enables one-shot decoding of polar codes. In the proposed scheme, rather than using the information bit vectors as labels for training the neural network (NN) through supervised learning as the conventional scheme did, the NN is trained to function as a bounded distance decoder by leveraging the generator matrix of polar codes through self-supervised learning. This approach eliminates the reliance on predefined labels, empowering the potential to train directly on the actual data within communication systems and thereby enhancing the applicability. Furthermore, computer simulations demonstrate that (i) the bit error rate (BER) and block error rate (BLER) performances of the proposed scheme can approach those of the maximum a posteriori (MAP) decoder for very short packets and (ii) the proposed NN decoder exhibits much superior generalization ability compared to the conventional one.

翻译：我们提出了一种基于无监督深度学习的译码方案，能够实现极化码的单次解码。在所提方案中，神经网络（NN）并非像传统方案那样通过监督学习使用信息比特向量作为标签进行训练，而是通过自监督学习利用极化码的生成矩阵将其训练为有界距离译码器。该方法消除了对预定义标签的依赖，使得能够直接在通信系统中的真实数据上进行训练，从而增强了适用性。此外，计算机仿真表明：(i) 所提方案的误比特率（BER）和误块率（BLER）性能在超短数据包场景下可接近最大后验概率（MAP）译码器；(ii) 与传统方案相比，所提神经网络译码器具有显著更强的泛化能力。