Challenges
Datasets
Workspaces
Papers
Discussions
Leaderboard
Log inSign up
Challenges
Datasets
Workspaces
Papers
Discussions
Leaderboard
Blog
Job Board
Q3AS

© 2026 Aqora Quantum S.A.S.

TermsPrivacyLegal Notice

IEEE Information Theory Workshop (ITW)

A Decoder with Reinforcement Learning Feedback

Milad Taghipour, Bane V. Vasic

Published Sep 29, 2025• 0 citations
Publisher

Abstract

This paper explores the application of reinforcement learning techniques to improve the performance of bit-flipping decoders and finding their optimal decisions. We begin by providing an overview of bit-flipping based decoders and reinforcement learning algorithms. We then outline the methodology for mapping the iterative decoding process into Markov Decision Processes (MDPs). We propose a feedback-based method to exploit and enhance the performance of existing well-established decoders by applying reinforcement learning algorithms after a selected decoder. In essence, we modify the MDP to reduce the number of states, making reinforcement learning algorithms feasible for low-rate and long-length codes. In addition, focusing on the correction of dominant error patterns of a selected decoder improves its error correction capability. Finally, we present experimental results for the Binary Symmetric Channel (BSC) to demonstrate the efficiency of the proposed methods.

DOI
10.1109/itw62417.2025.11240406

Related papers

Action-List Reinforcement Learning Decoders

Milad Taghipour, Bane V. Vasic

This paper explores the application of reinforcement learning techniques to enhance the performance of decoding based on flipping bits and finding optimal decisions. We begin by providing an overview of bit-flipping based decoders and reinforcement learning algorithms. We then describe the methodology for mapping the iterative decoding process into Markov Decision Processes (MDPs) and propose a general action-list decoding method for reinforcement learning based decoders irrespective of the class of codes to improve the performance of decoders. We design an action-list decoder based on the Deep-Q network values that substantially enhance performance. We also get benefit of automorphism group of code to further improve the code performance. Finally, we present experimental results for the Binary Symmetric Channel (BSC) to demonstrate the efficiency of the proposed methods.

International Symposium on Turbo Codes and Iterative Information Processing•Aug 18, 2025•0 citations

Reinforcement Learning for Channel Coding: Learned Bit-Flipping Decoding

Fabrizio Carpi, Christian Häger, Marco Martalò +2 more

In this paper, we use reinforcement learning to find effective decoding strategies for binary linear codes. We start by reviewing several iterative decoding algorithms that involve a decision-making process at each step, including bit-flipping (BF) decoding, residual belief propagation, and anchor decoding. We then illustrate how such algorithms can be mapped to Markov decision processes allowing for data-driven learning of optimal decision strategies, rather than basing decisions on heuristics or intuition. As a case study, we consider BF decoding for both the binary symmetric and additive white Gaussian noise channel. Our results show that learned BF decoders can offer a range of performance-complexity trade-offs for the considered Reed-Muller and BCH codes, and achieve near-optimal performance in some cases. We also demonstrate learning convergence speed-ups when biasing the learning process towards correct decoding decisions, as opposed to relying only on random explorations and past knowledge.

•Jan 1, 2019•20 citations

Reinforcement Learning for Channel Coding: Learned Bit-Flipping Decoding

Fabrizio Carpi, Christian Häger, Marco Martalò +2 more

In this paper, we use reinforcement learning to find effective decoding strategies for binary linear codes. We start by reviewing several iterative decoding algorithms that involve a decision-making process at each step, including bit-flipping (BF) decoding, residual belief propagation, and anchor decoding. We then illustrate how such algorithms can be mapped to Markov decision processes allowing for data-driven learning of optimal decision strategies, rather than basing decisions on heuristics or intuition. As a case study, we consider BF decoding for both the binary symmetric and additive white Gaussian noise channel. Our results show that learned BF decoders can offer a range of performance-complexity trade-offs for the considered Reed-Muller and BCH codes, and achieve near-optimal performance in some cases. We also demonstrate learning convergence speed-ups when biasing the learning process towards correct decoding decisions, as opposed to relying only on random explorations and past knowledge.

arXiv:1906.04448•Allerton Conference on Communication, Control, and Computing•Sep 1, 2019•44 citations

Reinforcement Learning for Enhanced Advanced QEC Architecture Decoding

Yidong Zhou, Lingyi Kong, Yifeng Peng +1 more

The advent of promising quantum error correction (QEC) codes with efficient resource utilization and high-performance fault-tolerant quantum memories signifies a critical step towards realizing practical quantum computation. While surface codes have been a dominant approach, their limitations have spurred the development of more advanced QEC architectures. These advanced codes often present increased complexity, demanding innovative decoding methodologies. This work investigates the application of reinforcement learning (RL) techniques, including hybrid and multi-agent approaches, to enhance the decoding of various advanced QEC architectures. By leveraging the ability of RL to learn optimal strategies from noisy syndrome measurements, we explore the potential for achieving improved logical error rates and scalability compared to traditional decoding methods. Our approach examines the adaptation of reinforcement learning to exploit the structural properties of these modern QEC models. We also explore the benefits of combining different RL algorithms to address the multifaceted nature of the decoding problem, considering factors such as code degeneracy and real-world noise characteristics. With our proposed method, we are able to demonstrate that an autonomously trained agent can derive decoding schemes for the complex decoding requirement of advanced QEC architectures.

quant-pharXiv:2601.19279•Jan 28, 2026•0 citations

Reinforcement Learning for Channel Coding

Mengke Lian, Fabrizio Carpi, Christian Häger +1 more

We discuss the application of reinforcement learning (RL) to problems associated with decoding binary linear codes.In particular, we consider two different iterative decoding algorithms that involve sequential decisions and apply RL to optimize those decisions.For the first example, we consider bit-flipping (BF) decoders and observe that learned BF decoders can offer a range of performance-complexity trade-offs and achieve nearoptimal performance in some cases.For the second example, we consider protograph low-density parity-check (LDPC) codes and use RL to optimize the decoding schedule.Our results show that, the comparison with the flooding schedule, a fixed error rate can be achieved with fewer update operations.

•Feb 26, 2020•0 citations

A Reinforcement Learning Based Decoding Method of Short Polar Codes

Jian Shu Gao, Kai Niu

In this paper, reinforcement learning (RL) is applied to find effective decoding strategies for short polar codes. We first illustrate how to map the step-decision decoding to a Markov decision process. Then, we choose the signal reliability to formulate the reward strategy, and the path metrics are taken as the Q-values of the state-action pairs. Finally, an adaptive Q-table approach is proposed for data-driven learning of optimal decision strategies. Compared with conventional Q-learning and deep Q-learning networks, the adaptive Q-table results in low computational and storage complexity. Simulation results show that the proposed RL decoding with adaptive Q-table achieves comparable performance to learned bit-flipping decoding and SC decoding with low complexity.

IEEE Wireless Communications and Networking Conference Workshops (WCNCW)•Mar 29, 2021•3 citations
This paper explores the application of reinforcement learning techniques to improve the performance of bit-flipping decoders and finding their optimal decisions. We begin by providing an overview of bit-flipping based decoders and reinforcement learning algorithms. We then outline the methodology for mapping the iterative decoding process into Markov Decision Processes (MDPs). We propose a feedback-based method to exploit and enhance the performance of existing well-established decoders by applying reinforcement learning algorithms after a selected decoder. In essence, we modify the MDP to reduce the number of states, making reinforcement learning algorithms feasible for low-rate and long-length codes. In addition, focusing on the correction of dominant error patterns of a selected decoder improves its error correction capability. Finally, we present experimental results for the Binary Symmetric Channel (BSC) to demonstrate the efficiency of the proposed methods.
This paper explores the application of reinforcement learning techniques to enhance the performance of decoding based on flipping bits and finding optimal decisions. We begin by providing an overview of bit-flipping based decoders and reinforcement learning algorithms. We then describe the methodology for mapping the iterative decoding process into Markov Decision Processes (MDPs) and propose a general action-list decoding method for reinforcement learning based decoders irrespective of the class of codes to improve the performance of decoders. We design an action-list decoder based on the Deep-Q network values that substantially enhance performance. We also get benefit of automorphism group of code to further improve the code performance. Finally, we present experimental results for the Binary Symmetric Channel (BSC) to demonstrate the efficiency of the proposed methods.
In this paper, we use reinforcement learning to find effective decoding strategies for binary linear codes. We start by reviewing several iterative decoding algorithms that involve a decision-making process at each step, including bit-flipping (BF) decoding, residual belief propagation, and anchor decoding. We then illustrate how such algorithms can be mapped to Markov decision processes allowing for data-driven learning of optimal decision strategies, rather than basing decisions on heuristics or intuition. As a case study, we consider BF decoding for both the binary symmetric and additive white Gaussian noise channel. Our results show that learned BF decoders can offer a range of performance-complexity trade-offs for the considered Reed-Muller and BCH codes, and achieve near-optimal performance in some cases. We also demonstrate learning convergence speed-ups when biasing the learning process towards correct decoding decisions, as opposed to relying only on random explorations and past knowledge.
In this paper, we use reinforcement learning to find effective decoding strategies for binary linear codes. We start by reviewing several iterative decoding algorithms that involve a decision-making process at each step, including bit-flipping (BF) decoding, residual belief propagation, and anchor decoding. We then illustrate how such algorithms can be mapped to Markov decision processes allowing for data-driven learning of optimal decision strategies, rather than basing decisions on heuristics or intuition. As a case study, we consider BF decoding for both the binary symmetric and additive white Gaussian noise channel. Our results show that learned BF decoders can offer a range of performance-complexity trade-offs for the considered Reed-Muller and BCH codes, and achieve near-optimal performance in some cases. We also demonstrate learning convergence speed-ups when biasing the learning process towards correct decoding decisions, as opposed to relying only on random explorations and past knowledge.
The advent of promising quantum error correction (QEC) codes with efficient resource utilization and high-performance fault-tolerant quantum memories signifies a critical step towards realizing practical quantum computation. While surface codes have been a dominant approach, their limitations have spurred the development of more advanced QEC architectures. These advanced codes often present increased complexity, demanding innovative decoding methodologies. This work investigates the application of reinforcement learning (RL) techniques, including hybrid and multi-agent approaches, to enhance the decoding of various advanced QEC architectures. By leveraging the ability of RL to learn optimal strategies from noisy syndrome measurements, we explore the potential for achieving improved logical error rates and scalability compared to traditional decoding methods. Our approach examines the adaptation of reinforcement learning to exploit the structural properties of these modern QEC models. We also explore the benefits of combining different RL algorithms to address the multifaceted nature of the decoding problem, considering factors such as code degeneracy and real-world noise characteristics. With our proposed method, we are able to demonstrate that an autonomously trained agent can derive decoding schemes for the complex decoding requirement of advanced QEC architectures.
We discuss the application of reinforcement learning (RL) to problems associated with decoding binary linear codes.In particular, we consider two different iterative decoding algorithms that involve sequential decisions and apply RL to optimize those decisions.For the first example, we consider bit-flipping (BF) decoders and observe that learned BF decoders can offer a range of performance-complexity trade-offs and achieve nearoptimal performance in some cases.For the second example, we consider protograph low-density parity-check (LDPC) codes and use RL to optimize the decoding schedule.Our results show that, the comparison with the flooding schedule, a fixed error rate can be achieved with fewer update operations.
In this paper, reinforcement learning (RL) is applied to find effective decoding strategies for short polar codes. We first illustrate how to map the step-decision decoding to a Markov decision process. Then, we choose the signal reliability to formulate the reward strategy, and the path metrics are taken as the Q-values of the state-action pairs. Finally, an adaptive Q-table approach is proposed for data-driven learning of optimal decision strategies. Compared with conventional Q-learning and deep Q-learning networks, the adaptive Q-table results in low computational and storage complexity. Simulation results show that the proposed RL decoding with adaptive Q-table achieves comparable performance to learned bit-flipping decoding and SC decoding with low complexity.