Academic Bibliography

 Search 200 years of publications by Ghent University researchers.
Advanced search
1 file | 4.41 MB
Add to list
  1. Search results
  2. Programmable tanh- and ELU-based phot...

Programmable tanh- and ELU-based photonic neurons in optics-informed neural networks

Stefanos Kovaios, Christos Pappas, Miltiadis Moralis-Pegios, Apostolos Tsakyridis, George Giamougiannis, Manos Kirtas, Joris Van Kerrebrouck (UGent) , Gertjan Coudyzer (UGent) , Xin Yin (UGent) , Nikolaos Passalis, et al.
(2024) JOURNAL OF LIGHTWAVE TECHNOLOGY. 42(10). p.3652-3660
Author
Organization
Abstract
We demonstrate an integrated opto-electronic (Omicron Epsilon) device that can be programmed to provide a set of nonlinear activation functions (AFs) and present its operation within programmable tanh- and ELU-based photonic neurons at line rates up to 10 GBd. The OE activation module provides a set of well-known activation functions that are typically used in DL training models, including the tanh-, ELU- and inverted ELU-like functions. Its performance is experimentally evaluated when incorporated in a 4-input wavelength division multiplexed (WDM) photonic neuron and operating with non-deterministic data patterns, providing "noisy" tanh, ELU and inverted ELU AFs with an error-distribution that has in all cases a standard deviation of <0.49. We also evaluate the trainability of these "noisy" AFs and present for the first time an optics-informed training framework that incorporates the pattern-induced AF variations into the training process, yielding the first noise-aware training scheme where the noise emerges at the nonlinear AF NN segment. The performance analysis of the optics-informed training framework for all three AFs was carried out via Deep Learning setups suitable for classifying the Fashion MNIST and the CIFAR-10 datasets. This analysis has shown that the employment of traditional training schemes leads to significant accuracy degradations, which can be, however, almost completely waived when employing the optics-informed training framework, leading to accuracy values that are almost identical to the reference accuracy values obtained when ideal and noise-less AFs are used.
Keywords
Nonlinear optics, Photonics, Neurons, Training, Integrated circuits, Artificial neural networks, Adaptive optics, Noise aware training, nonlinear activation function, optical neural networks, opto-electronic activation function, ACTIVATION FUNCTIONS, ARTIFICIAL-INTELLIGENCE

Downloads

  • Download
    (...).pdf
    • full text (Published version)
    • |
    • UGent only
    • |
    • PDF
    • |
    • 4.41 MB

Citation

Please use this url to cite or link to this publication:

MLA
Kovaios, Stefanos, et al. “Programmable Tanh- and ELU-Based Photonic Neurons in Optics-Informed Neural Networks.” JOURNAL OF LIGHTWAVE TECHNOLOGY, vol. 42, no. 10, 2024, pp. 3652–60, doi:10.1109/JLT.2024.3366711.
APA
Kovaios, S., Pappas, C., Moralis-Pegios, M., Tsakyridis, A., Giamougiannis, G., Kirtas, M., … Pleros, N. (2024). Programmable tanh- and ELU-based photonic neurons in optics-informed neural networks. JOURNAL OF LIGHTWAVE TECHNOLOGY, 42(10), 3652–3660. https://doi.org/10.1109/JLT.2024.3366711
Chicago author-date
Kovaios, Stefanos, Christos Pappas, Miltiadis Moralis-Pegios, Apostolos Tsakyridis, George Giamougiannis, Manos Kirtas, Joris Van Kerrebrouck, et al. 2024. “Programmable Tanh- and ELU-Based Photonic Neurons in Optics-Informed Neural Networks.” JOURNAL OF LIGHTWAVE TECHNOLOGY 42 (10): 3652–60. https://doi.org/10.1109/JLT.2024.3366711.
Chicago author-date (all authors)
Kovaios, Stefanos, Christos Pappas, Miltiadis Moralis-Pegios, Apostolos Tsakyridis, George Giamougiannis, Manos Kirtas, Joris Van Kerrebrouck, Gertjan Coudyzer, Xin Yin, Nikolaos Passalis, Anastasios Tefas, and Nikos Pleros. 2024. “Programmable Tanh- and ELU-Based Photonic Neurons in Optics-Informed Neural Networks.” JOURNAL OF LIGHTWAVE TECHNOLOGY 42 (10): 3652–3660. doi:10.1109/JLT.2024.3366711.
Vancouver
1.
Kovaios S, Pappas C, Moralis-Pegios M, Tsakyridis A, Giamougiannis G, Kirtas M, et al. Programmable tanh- and ELU-based photonic neurons in optics-informed neural networks. JOURNAL OF LIGHTWAVE TECHNOLOGY. 2024;42(10):3652–60.
IEEE
[1]
S. Kovaios et al., “Programmable tanh- and ELU-based photonic neurons in optics-informed neural networks,” JOURNAL OF LIGHTWAVE TECHNOLOGY, vol. 42, no. 10, pp. 3652–3660, 2024.
@article{01HYD82NZKBN034PX5WAWA6R6C,
  abstract     = {{We demonstrate an integrated opto-electronic (Omicron Epsilon) device that can be programmed to provide a set of nonlinear activation functions (AFs) and present its operation within programmable tanh- and ELU-based photonic neurons at line rates up to 10 GBd. The OE activation module provides a set of well-known activation functions that are typically used in DL training models, including the tanh-, ELU- and inverted ELU-like functions. Its performance is experimentally evaluated when incorporated in a 4-input wavelength division multiplexed (WDM) photonic neuron and operating with non-deterministic data patterns, providing "noisy" tanh, ELU and inverted ELU AFs with an error-distribution that has in all cases a standard deviation of <0.49. We also evaluate the trainability of these "noisy" AFs and present for the first time an optics-informed training framework that incorporates the pattern-induced AF variations into the training process, yielding the first noise-aware training scheme where the noise emerges at the nonlinear AF NN segment. The performance analysis of the optics-informed training framework for all three AFs was carried out via Deep Learning setups suitable for classifying the Fashion MNIST and the CIFAR-10 datasets. This analysis has shown that the employment of traditional training schemes leads to significant accuracy degradations, which can be, however, almost completely waived when employing the optics-informed training framework, leading to accuracy values that are almost identical to the reference accuracy values obtained when ideal and noise-less AFs are used.}},
  author       = {{Kovaios, Stefanos and  Pappas, Christos and  Moralis-Pegios, Miltiadis and  Tsakyridis, Apostolos and  Giamougiannis, George and  Kirtas, Manos and Van Kerrebrouck, Joris and Coudyzer, Gertjan and Yin, Xin and  Passalis, Nikolaos and  Tefas, Anastasios and  Pleros, Nikos}},
  issn         = {{0733-8724}},
  journal      = {{JOURNAL OF LIGHTWAVE TECHNOLOGY}},
  keywords     = {{Nonlinear optics,Photonics,Neurons,Training,Integrated circuits,Artificial neural networks,Adaptive optics,Noise aware training,nonlinear activation function,optical neural networks,opto-electronic activation function,ACTIVATION FUNCTIONS,ARTIFICIAL-INTELLIGENCE}},
  language     = {{eng}},
  number       = {{10}},
  pages        = {{3652--3660}},
  title        = {{Programmable tanh- and ELU-based photonic neurons in optics-informed neural networks}},
  url          = {{http://doi.org/10.1109/JLT.2024.3366711}},
  volume       = {{42}},
  year         = {{2024}},
}
Altmetric
View in Altmetric
Web of Science
Times cited: