DEEPTURBO: Deep Turbo Decoder

Abstract

Present-day communication systems routinely use codes that approach the channel capacity when coupled with a computationally efficient decoder. However, the decoder is typically designed for the Gaussian noise channel, and is known to be sub-optimal for non-Gaussian noise distribution. Deep learning methods offer a new approach for designing decoders that can be trained and tailored for arbitrary channel statistics. We focus on Turbo codes, and propose (DEEPTURBO), a novel deep learning based architecture for Turbo decoding.The standard Turbo decoder (TURBO) iteratively applies the Bahl-Cocke-Jelinek-Raviv (BCJR) algorithm with an interleaver in the middle. A neural architecture for Turbo decoding, termed (NEURALBCJR), was proposed recently. There, the key idea is to create a module that imitates the BCJR algorithm using supervised learning, and to use the interleaver architecture along with this module, which is then fine-tuned using end-to-end training. However, knowledge of the BCJR algorithm is required to design such an architecture, which also constrains the resulting learnt decoder. Here we remedy this requirement and propose a fully end-to-end trained neural decoder -Deep Turbo Decoder (DEEPTURBO). With novel learnable decoder structure and training methodology, DEEPTURBO reveals superior performance under both AWGN and non-AWGN settings as compared to the other two decoders -TURBO and NEURALBCJR. Furthermore, among all the three, DEEPTURBO exhibits the lowest error floor.

Active Deep Decoding of Linear Codes

“…Utilizing concurrent NN designs in addition with learning the code properties, via the composed syndrome, achieved performance improvement. Further contributions to the field lie in [10] and [11], where neural decoders for convolutional codes were proposed together with befitting training methodologies. However, these methodologies impose substantial increase of complexity, at both training and decoding (inference).…”

“…However, these methodologies impose substantial increase of complexity, at both training and decoding (inference). Specifically, [10] explores a recurrent neural networks (RNN) architecture for decoding while [11] focuses on an unconstrained novel structure requiring no knowledge of the BCJR (Bahl-Cocke-Jelinek-Raviv) algorithm.…”

On Recurrent Neural Networks for Sequence-based Processing in Communications

“…The huge success of deep learning (DL) and neural networks (NNs), mainly in the fields of computer vision and speech processing, has recently triggered further exploration of DL for communications. The potential applications span from trainable channel decoders [1], [2], [3], [4], [5], NNbased multiple-input multiple-output (MIMO) detectors [6] and detectors for molecular channels [7] up to communication systems that inherently learn to communicate [8]. Most of these applications typically rely on block-based signal processing, which is an obvious consequence if NN structures are used that were derived from computer vision tasks.…”

“…The universal approximator theorem [18], and the fact that an explicit optimal algorithm exists, directly tells us that also an NN must exist (neglecting any complexity constraints) which comes arbitrarily close to the optimal performance. While other groups already showed the existence of such decoding NNs for short memory convolutional codes [2], [3], we want to further investigate to what extent NNs are capable of processing even more complex information sequences using the example of decoding convolutional codes up to memory ν. However, in practice the limiting factor is clearly the training complexity and, thus, the major challenge is to find a suitable training method for this task.…”

DeepJSCC-f: Deep Joint Source-Channel Coding of Images With Feedback

“…The decoder of any known channel code can be treated as a classifier on the noisy channel output. With the growing success of deep learning for various classification tasks, researchers first explored using DNNs to decode existing channel codes, such as linear [23], turbo [24], or polar [25] codes, which provided promising improvements in the error probability. This was later extended to the joint design of the encoder and the decoder using an autoencoder structure [26].…”