4.7 Article

Intra-Channel Nonlinearity Mitigation in Optical Fiber Transmission Systems Using Perturbation-Based Neural Network

Related references

Note: Only part of the references are listed.
Article Computer Science, Artificial Intelligence

Distilled Neural Networks for Efficient Learning to Rank

Franco Maria Nardini et al.

Summary: Recent studies have shown the possibility of distilling a neural network from an ensemble of regression trees, making neural networks a competitor to tree-based ensembles in ranking tasks. However, regression tree ensembles outperform neural models in efficiency and effectiveness when scoring on CPU. In this paper, we propose a method to speed up neural scoring time by combining Distillation, Pruning, and Fast Matrix multiplication. We introduce knowledge distillation to learn shallow neural networks from regression tree ensembles and employ an efficiency-oriented pruning technique to sparsify computationally-intensive layers, scoring them with optimized sparse matrix multiplication. We also develop a scoring time prediction model by studying dense and sparse high performance matrix multiplication, aiding in the design of efficient neural network architectures. Comprehensive experiments on public learning-to-rank datasets demonstrate the competitiveness of our approach, providing up to 4x scoring time speed-up without compromising ranking quality.

IEEE TRANSACTIONS ON KNOWLEDGE AND DATA ENGINEERING (2023)

Article Engineering, Electrical & Electronic

Neural Networks-Based Equalizers for Coherent Optical Transmission: Caveats and Pitfalls

Pedro J. Freire et al.

Summary: This paper provides a detailed and multi-faceted analysis of the challenges and design issues in the development of efficient neural network-based nonlinear channel equalizers in coherent optical communication systems. The study clarifies the evaluation metrics for equalizer performance, establishes the relationship between channel propagation model accuracy and equalizer performance, and examines the impact of pseudo-random bit sequence order and DAC memory limitations on equalizer operation. The paper also addresses overfitting limitations, classification versus regression, and batch-size-related peculiarities in NN equalizers.

IEEE JOURNAL OF SELECTED TOPICS IN QUANTUM ELECTRONICS (2022)

Article Engineering, Electrical & Electronic

Physics-Based Deep Learning for Fiber-Optic Communication Systems

Christian Hager et al.

Summary: A new physics-based machine-learning model is proposed for solving problems in fiber-optic communication systems, efficiently inverting the nonlinear Schrodinger equation. By progressively pruning filter taps, complexity can be significantly reduced.

IEEE JOURNAL ON SELECTED AREAS IN COMMUNICATIONS (2021)

Article Engineering, Electrical & Electronic

End-to-End Optimization of Coherent Optical Communications Over the Split-Step Fourier Method Guided by the Nonlinear Fourier Transform Theory

Simone Gaiarin et al.

Summary: Optimizing modulation and detection strategies is crucial for maximizing communication system throughput, which can be easily carried out analytically for linear channels but becomes challenging for nonlinear channels such as optical fibers. A novel autoencoder scheme applied to the full optical channel described by the nonlinear Schrodinger equation is proposed in this study, demonstrating a significant improvement in system performance.

JOURNAL OF LIGHTWAVE TECHNOLOGY (2021)

Article Engineering, Electrical & Electronic

Performance Versus Complexity Study of Neural Network Equalizers in Coherent Optical Systems

Pedro J. Freire et al.

Summary: The study compared the performance and complexity of different artificial neural networks for nonlinear channel equalization in coherent optical communication systems. The CNN+biLSTM architecture showed the largest Q-factor improvement and was the best option when computational complexity was not constrained. However, when complexity was restricted to lower levels, the three-layer perceptron performed best.

JOURNAL OF LIGHTWAVE TECHNOLOGY (2021)

Article Engineering, Electrical & Electronic

A Fiber Nonlinearity Compensation Scheme With Complex-Valued Dimension-Reduced Neural Network

Pinjing He et al.

Summary: The proposed fiber nonlinearity compensation scheme based on a complex-valued dimension-reduced neural network showed higher efficiency in computational complexity compared to the real-valued input neural network, as well as better performance in noise robustness and weight quantization.

IEEE PHOTONICS JOURNAL (2021)

Article Engineering, Electrical & Electronic

Neural Network Nonlinear Equalizer in Long-Distance Coherent Optical Transmission Systems

T. Kamiyama et al.

Summary: This study proposes a method to train neural networks for compensating fiber nonlinearities in long-distance coherent optical transmission systems. The method involves estimating phase noise by inverse modulation of received signals and using it for training. Results show the effectiveness of this technique in the presence of phase noise.

IEEE PHOTONICS TECHNOLOGY LETTERS (2021)

Article Engineering, Electrical & Electronic

Complex-Valued Neural Network Design for Mitigation of Signal Distortions in Optical Links

Pedro J. Freire et al.

Summary: Nonlinearity compensation is crucial for increasing channel transmission rates in optical communication systems. Data-driven approaches using neural networks have shown to improve the performance of complex fiber-optic systems without prior knowledge of specific parameters. The proposed neural network design, optimized using a Bayesian optimizer, has successfully demonstrated improved performance in both linear and nonlinear regimes of fiber-optic communication.

JOURNAL OF LIGHTWAVE TECHNOLOGY (2021)

Article Engineering, Electrical & Electronic

Advanced Convolutional Neural Networks for Nonlinearity Mitigation in Long-Haul WDM Transmission Systems

Oleg Sidelnikov et al.

Summary: The study utilizes convolutional neural networks to compensate for nonlinear signal distortions in optical communication systems, optimizing algorithmic complexity to improve learning efficiency and performance.

JOURNAL OF LIGHTWAVE TECHNOLOGY (2021)

Article Optics

Information rates in Kerr nonlinearity limited optical fiber communication systems

Tianhua Xu et al.

Summary: The achievable information rates of optical communication systems are limited by nonlinear distortions due to the Kerr effect in optical fibers, with these impairments becoming more significant for systems with larger transmission bandwidths, closer channel spacing, and higher-order modulation formats. This paper investigates the effectiveness of nonlinearity compensation techniques in enhancing achievable information rates in fully-loaded C-band systems, considering practical transceiver limitations. The study shows that the performance of nonlinearity compensation techniques strongly depends on signal modulation formats and target transmission distances in C-band systems.

OPTICS EXPRESS (2021)

Article Engineering, Electrical & Electronic

Nonlinear Coherent Optical Systems in the Presence of Equalization Enhanced Phase Noise

Cenqin Jin et al.

Summary: The paper investigates the impact of Equalization Enhanced Phase Noise (EEPN) on multi-channel optical communication systems, analyzes the nonlinear interaction between the signal and EEPN, and validates the analytical approach through numerical simulations and split-step Fourier simulations.

JOURNAL OF LIGHTWAVE TECHNOLOGY (2021)

Article Engineering, Electrical & Electronic

Performance and Complexity Analysis of Bi-Directional Recurrent Neural Network Models Versus Volterra Nonlinear Equalizers in Digital Coherent Systems

Stavros Deligiannidis et al.

Summary: The study investigates the complexity and performance of RNN models as post-processing units for compensating fiber nonlinearities in digital coherent systems. The results show that Vanilla-RNN units are the preferred choice due to their simplicity. Bi-directional Vanilla-RNN outperforms Volterra nonlinear equalizers in terms of both performance and complexity, indicating that RNN processing is a promising pathway for upgrading long-haul optical communication systems.

JOURNAL OF LIGHTWAVE TECHNOLOGY (2021)

Article Computer Science, Artificial Intelligence

Binary neural networks: A survey

Haotong Qin et al.

PATTERN RECOGNITION (2020)

Article Engineering, Electrical & Electronic

Compensation of Nonlinear Impairments Using Inverse Perturbation Theory With Reduced Complexity

Alexey Redyuk et al.

JOURNAL OF LIGHTWAVE TECHNOLOGY (2020)

Article Computer Science, Information Systems

Overfitting effect of artificial neural network based nonlinear equalizer: from mathematical origin to transmission evolution

Zheng Yang et al.

SCIENCE CHINA-INFORMATION SCIENCES (2020)

Review Computer Science, Hardware & Architecture

Machine learning for intelligent optical networks: A comprehensive survey

Rentao Gu et al.

JOURNAL OF NETWORK AND COMPUTER APPLICATIONS (2020)

Article Engineering, Electrical & Electronic

Data-driven Optical Fiber Channel Modeling: A Deep Learning Approach

Danshi Wang et al.

JOURNAL OF LIGHTWAVE TECHNOLOGY (2020)

Article Engineering, Electrical & Electronic

Enabling Technologies for Fiber Nonlinearity Mitigation in High Capacity Transmission Systems

Olga Vassilieva et al.

JOURNAL OF LIGHTWAVE TECHNOLOGY (2019)

Article Engineering, Electrical & Electronic

Perturbation-Based FEC-Assisted Iterative Nonlinearity Compensation for WDM Systems

Edson Porto da Silva et al.

JOURNAL OF LIGHTWAVE TECHNOLOGY (2019)

Article Multidisciplinary Sciences

Field and lab experimental demonstration of nonlinear impairment compensation using neural networks

Shaoliang Zhang et al.

NATURE COMMUNICATIONS (2019)

Article Engineering, Electrical & Electronic

Achievable Information Rates for Fiber Optics: Applications and Computations

Alex Alvarado et al.

JOURNAL OF LIGHTWAVE TECHNOLOGY (2018)

Article Engineering, Electrical & Electronic

Fiber Nonlinearity Mitigation in Single Carrier 400 G and 800 G Nyquist-WDM Systems

Chenjia Li et al.

JOURNAL OF LIGHTWAVE TECHNOLOGY (2018)

Article Engineering, Electrical & Electronic

Applying Neural Networks in Optical Communication Systems: Possible Pitfalls

Tobias A. Eriksson et al.

IEEE PHOTONICS TECHNOLOGY LETTERS (2017)

Article Engineering, Electrical & Electronic

Digital Signal Processing for Coherent Transceivers Employing Multilevel Formats

Md. Saifuddin Faruk et al.

JOURNAL OF LIGHTWAVE TECHNOLOGY (2017)

Article Optics

Performance limits in optical communications due to fiber nonlinearity

A. D. Ellis et al.

ADVANCES IN OPTICS AND PHOTONICS (2017)

Article Engineering, Electrical & Electronic

Machine Learning Techniques in Optical Communication

Darko Zibar et al.

JOURNAL OF LIGHTWAVE TECHNOLOGY (2016)

Article Engineering, Electrical & Electronic

Fiber Nonlinearity Mitigation in 32-Gbaud 16QAM Nyquist-WDM Systems

Zhennan Zheng et al.

JOURNAL OF LIGHTWAVE TECHNOLOGY (2016)

Article Multidisciplinary Sciences

Maximizing the optical network capacity

Polina Bayvel et al.

PHILOSOPHICAL TRANSACTIONS OF THE ROYAL SOCIETY A-MATHEMATICAL PHYSICAL AND ENGINEERING SCIENCES (2016)

Article Engineering, Electrical & Electronic

Fiber Nonlinearity Equalizer Based on Support Vector Classification for Coherent Optical OFDM

Tu Nguyen et al.

IEEE PHOTONICS JOURNAL (2016)

Review Multidisciplinary Sciences

Deep learning

Yann LeCun et al.

NATURE (2015)

Article Optics

On achievable rates for long-haul fiber-optic communications

Tobias Fehenberger et al.

OPTICS EXPRESS (2015)

Article Engineering, Electrical & Electronic

Advanced DSP for 400 Gb/s and Beyond Optical Networks

Xiang Zhou et al.

JOURNAL OF LIGHTWAVE TECHNOLOGY (2014)

Article Optics

EGN model of non-linear fiber propagation

Andrea Carena et al.

OPTICS EXPRESS (2014)

Article Engineering, Electrical & Electronic

Nonlinearity Compensation of 224 Gb/s Dual-Polarization 16-QAM Transmission Over 2700 km

Ying Gao et al.

IEEE PHOTONICS TECHNOLOGY LETTERS (2013)

Article Engineering, Electrical & Electronic

Achievable Information Rate in Nonlinear WDM Fiber-Optic Systems With Arbitrary Modulation Formats and Dispersion Maps

Marco Secondini et al.

JOURNAL OF LIGHTWAVE TECHNOLOGY (2013)

Article Engineering, Electrical & Electronic

Nonlinear Shannon Limit in Pseudolinear Coherent Systems

Antonio Mecozzi et al.

JOURNAL OF LIGHTWAVE TECHNOLOGY (2012)

Article Engineering, Electrical & Electronic

Multiplier-Free Intrachannel Nonlinearity Compensating Algorithm Operating at Symbol Rate

Zhenning Tao et al.

JOURNAL OF LIGHTWAVE TECHNOLOGY (2011)

Article Engineering, Electrical & Electronic

Capacity Limits of Optical Fiber Networks

Rene-Jean Essiambre et al.

JOURNAL OF LIGHTWAVE TECHNOLOGY (2010)

Review Optics

Coherent detection in optical fiber systems

Ezra Ip et al.

OPTICS EXPRESS (2008)

Article Computer Science, Information Systems

Simulation-based computation of information rates for channels with memory

Dieter M. Arnold et al.

IEEE TRANSACTIONS ON INFORMATION THEORY (2006)

Article Engineering, Electrical & Electronic

System impact of intra-channel nonlinear effects in highly dispersed optical pulse transmission

A Mecozzi et al.

IEEE PHOTONICS TECHNOLOGY LETTERS (2000)