@ARTICLE{8433895,
  author={B. {Karanov} and M. {Chagnon} and F. {Thouin} and T. A. {Eriksson} and H. {Bülow} and D. {Lavery} and P. {Bayvel} and L. {Schmalen}},
  journal={Journal of Lightwave Technology}, 
  title={End-to-End Deep Learning of Optical Fiber Communications}, 
  year={2018},
  volume={36},
  number={20},
  pages={4843-4855},
  abstract={In this paper, we implement an optical fiber communication system as an end-to-end deep neural network, including the complete chain of transmitter, channel model, and receiver. This approach enables the optimization of the transceiver in a single end-to-end process. We illustrate the benefits of this method by applying it to intensity modulation/direct detection (IM/DD) systems and show that we can achieve bit error rates below the 6.7% hard-decision forward error correction (HD-FEC) threshold. We model all componentry of the transmitter and receiver, as well as the fiber channel, and apply deep learning to find transmitter and receiver configurations minimizing the symbol error rate. We propose and verify in simulations a training method that yields robust and flexible transceivers that allow-without reconfiguration-reliable transmission over a large range of link dispersions. The results from end-to-end deep learning are successfully verified for the first time in an experiment. In particular, we achieve information rates of 42 Gb/s below the HD-FEC threshold at distances beyond 40 km. We find that our results outperform conventional IM/DD solutions based on two- and four-level pulse amplitude modulation with feedforward equalization at the receiver. Our study is the first step toward end-to-end deep learning based optimization of optical fiber communication systems.},
  keywords={error statistics;forward error correction;intensity modulation;learning (artificial intelligence);neural nets;optical fibre communication;optical fibre dispersion;optical modulation;pulse amplitude modulation;intensity modulation/direct detection systems;bit error rates;fiber channel;symbol error rate;end-to-end deep learning based optimization;optical fiber communication system;optical fiber communications;forward error correction;single end-to-end process;end-to-end deep neural network;size 40.0 km;Training;Machine learning;Receivers;Optical transmitters;Transceivers;Optimization;Communication systems;Deep learning;detection;machine learning;modulation;neural networks;optical fiber communication},
  doi={10.1109/JLT.2018.2865109},
  ISSN={1558-2213},
  month={Oct},}

@ARTICLE{8664650,
  author={B. {Zhu} and J. {Wang} and L. {He} and J. {Song}},
  journal={IEEE Journal on Selected Areas in Communications}, 
  title={Joint Transceiver Optimization for Wireless Communication PHY Using Neural Network}, 
  year={2019},
  volume={37},
  number={6},
  pages={1364-1373},
  abstract={Deep learning has a wide application in the area of natural language processing and image processing due to its strong ability of generalization. In this paper, we propose a novel neural network structure for jointly optimizing the transmitter and receiver in communication physical layer under fading channels. We build up a convolutional autoencoder to simultaneously conduct the role of modulation, equalization, and demodulation. The proposed system is able to design different mapping scheme from input bit sequences of arbitrary length to constellation symbols according to different channel environments. The simulation results show that the performance of neural network-based system is superior to traditional modulation and equalization methods in terms of time complexity and bit error rate under fading channels. The proposed system can also be combined with other coding techniques to further improve the performance. Furthermore, the proposed system network is more robust to channel variation than traditional communication methods.},
  keywords={computational complexity;error statistics;fading channels;learning (artificial intelligence);natural language processing;neural nets;radio transceivers;wireless communication PHY;deep learning;natural language processing;image processing;communication physical layer;fading channels;convolutional autoencoder;arbitrary length;constellation symbols;neural network-based system;traditional modulation;time complexity;bit error rate;system network;traditional communication methods;transceiver optimization;neural network structure;equalization methods;mapping scheme;channel environments;coding techniques;Receivers;Convolutional codes;Communication systems;Deep learning;Neurons;Modulation;Transmitters;Deep learning;modulation;equalization;autoencoder;frequency selective fading},
  doi={10.1109/JSAC.2019.2904361},
  ISSN={1558-0008},
  month={June},}

@ARTICLE{6975096,
  author={M. A. {Jarajreh} and E. {Giacoumidis} and I. {Aldaya} and S. T. {Le} and A. {Tsokanos} and Z. {Ghassemlooy} and N. J. {Doran}},
  journal={IEEE Photonics Technology Letters}, 
  title={Artificial Neural Network Nonlinear Equalizer for Coherent Optical OFDM}, 
  year={2015},
  volume={27},
  number={4},
  pages={387-390},
  abstract={We propose a novel low-complexity artificial neural network (ANN)-based nonlinear equalizer (NLE) for coherent optical orthogonal frequency-division multiplexing (CO-OFDM) and compare it with the recent inverse Volterra-series transfer function (IVSTF)-based NLE over up to 1000 km of uncompensated links. Demonstration of ANN-NLE at 80-Gb/s CO-OFDM using 16-quadrature amplitude modulation reveals a Q -factor improvement after 1000-km transmission of 3 and 1 dB with respect to the linear equalization and IVSTF-NLE, respectively.},
  keywords={neural nets;OFDM modulation;quadrature amplitude modulation;novel low-complexity artificial neural network-based nonlinear equalizer;novel low-complexity ANN-based NLE;CO-OFDM;coherent optical orthogonal frequency-division multiplexing;16-quadrature amplitude modulation;Q -factor improvement;IVSTF-NLE;OFDM;Artificial neural networks;Q-factor;Optical fibers;Nonlinear optics;Equalizers;Optical fiber networks;Optical communication;coherent optical fiber transmission;functional link artificial neural networks;nonlinear equalizer;OFDM;Optical communication;coherent optical fiber transmission;functional link artificial neural networks;nonlinear equalizer;OFDM},
  doi={10.1109/LPT.2014.2375960},
  ISSN={1941-0174},
  month={Feb},}

@INPROCEEDINGS{7011454,
  author={P. {Dondon} and J. {Carvalho} and R. {Gardere} and P. {Lahalle} and G. {Tsenov} and V. {Mladenov}},
  booktitle={12th Symposium on Neural Network Applications in Electrical Engineering (NEUREL)}, 
  title={Implementation of a feed-forward Artificial Neural Network in VHDL on FPGA}, 
  year={2014},
  volume={},
  number={},
  pages={37-40},
  abstract={Describing an Artificial Neural Network (ANN) using VHDL allows a further implementation of such a system on FPGA. Indeed, the principal point of using FPGA for ANNs is flexibility that gives it an advantage toward other systems like ASICS which are entirely dedicated to one unique architecture and allowance to parallel programming, which is inherent to ANN calculation system and one of their advantages. Usually FPGAs do not have unlimited logical resources integrated in a single package and this limitation forcesrequirement for optimizations for the design in order to have the best efficiency in terms of speed and resource consumption. This paper deals with the VHDL designing problems which can be encountered when trying to describe and implement such ANNs on FPGAs.},
  keywords={feedforward neural nets;field programmable gate arrays;hardware description languages;parallel programming;feedforward artificial neural network;FPGA;ASICS;parallel programming;ANN calculation system;logical resources;resource consumption;VHDL designing problems;Neurons;Field programmable gate arrays;Biological neural networks;Artificial neural networks;Random access memory;Read only memory;MATLAB;FPGA implementation;neural networks;nonlinear systems;VHDL},
  doi={10.1109/NEUREL.2014.7011454},
  ISSN={},
  month={Nov},}

@INPROCEEDINGS{5328349,
  author={V. {Gupta} and K. {Khare} and R. P. {Singh}},
  booktitle={2009 International Conference on Advances in Recent Technologies in Communication and Computing}, 
  title={FPGA Design and Implementation Issues of Artificial Neural Network Based PID Controllers}, 
  year={2009},
  volume={},
  number={},
  pages={860-862},
  abstract={This paper discusses implementation issues of FPGA and ANN based PID controllers. FPGA-based reconfigurable computing architectures are suitable for hardware implementation of neural networks. FPGA realization of ANNs with a large number of neurons is still a challenging task. This paper discusses the issues involved in implementation of a multi-input neuron with linear/nonlinear excitation functions using FPGA. It also suggests advantages of error self-recurrent neural networks over back propagation neural network.},
  keywords={backpropagation;field programmable gate arrays;integrated circuit design;neural nets;reconfigurable architectures;three-term control;FPGA design;artificial neural network;PID controller;reconfigurable computing architecture;multiinput neuron;nonlinear excitation function;error self-recurrent neural network;backpropagation neural network;Field programmable gate arrays;Artificial neural networks;Three-term control;Neurons;Control systems;Multi-layer neural network;Communication system control;Hardware;Neural networks;Computer networks;FPGA;PID controller;artificial neural networks;error self-recurrent neural networks},
  doi={10.1109/ARTCom.2009.182},
  ISSN={},
  month={Oct},}