Abstract
Neural network decoders (NNDs) for rate-compatible polar codes are studied in this paper. We consider a family of rate-compatible polar codes which are constructed from a single polar coding sequence as defined by 5G new radios. We propose a transfer learning technique for training multiple NNDs of the rate-compatible polar codes utilizing their inclusion property. The trained NND for a low rate code is taken as the initial state of NND training for the next smallest rate code. The proposed method provides quicker training as compared to separate learning of the NNDs according to numerical results. We additionally show that an underfitting problem of NND training due to low model complexity can be solved by transfer learning techniques.
Highlights
Polar codes, proposed by Arikan in [1], are the first error correcting codes to provably achieve the symmetric capacity with low complexity in binary-input discrete memoryless channels (B-DMCs).This result means that Shannon’s random codes, which achieve channel capacity, are replaced by a practical code with a low-complexity decoding algorithm [1]
We show that an underfitting problem of Neural network decoders (NNDs) training due to low model complexity can be solved by transfer learning techniques
We proposed a method of training NNDs for a family of rate-compatible polar codes
Summary
Polar codes, proposed by Arikan in [1], are the first error correcting codes to provably achieve the symmetric capacity with low complexity in binary-input discrete memoryless channels (B-DMCs). Entropy 2020, 22, 496 the advanced structures of DNN [6] Such small neural network decoders (NNDs) are used to form a decoder for longer polar codes in combination with BP processing [7]. Because the supervised learning method [5,6] trains a decoder with data from a specific code and channel, a straightforward approach is to train all the decoders separately. It is shown that transfer learning from low to high-rate codes can train a high-rate code decoder better than the conventional training methods.
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
