Reducing Training Time of Deep Learning Based Digital Backpropagation by Stacking

Bitachon, Bertold Ian; Eppenberger, Marco; Baeuerle, Benedikt; Leuthold, Juerg

doi:10.1109/lpt.2022.3162157

Cited by 4 publications

(1 citation statement)

References 14 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…An ANN's core benefit is the ability to construct a system model from the data at hand. It utilizes the backpropagation technique to extract texture characteristics in classifying images [4]. Meanwhile, the backpropagation approach is one of the supervised ANN methods examining the error for all neurons after processing a dataset.…”

Section: Introductionmentioning

confidence: 99%

Applying Artificial Neural Network Based on Backpropagation Method for Indonesian Sign Language Recognition

Irvanizam,

Horatius,

Sofyan

2023

IJCDS

View full text Add to dashboard Cite

Sign-language is a nonverbal language utilized by hearing loss and speech disorder people to interact and speak with each other. Sometimes people call it a visual-language utilizing hand gestures and changes in hand shape to express intentions and thoughts. In Indonesia, people use two sign language types commonly. They are Indonesian Signal System (ISS) and Indonesian Sign Language (ISL). In this paper, we investigated research on the recognition of ISL alphabets by proposing an application using the Artificial Neural Network algorithm with the principle of the Backpropagation method for dealing with this research. Firstly, we performed preprocessing the image data sequentially by splitting the dataset, labeling the dataset based on the alphabet types, scaling the images into a size of 150x150 pixels, and segmentizing the datasets to eliminate noise in the image to yield color, grayscale, binary segmentations, and image edge detection. Later on, we split the dataset and distributed them into training and testing sets with two different proportions, 70:30 and 80:20 percent. After applying the algorithm, we obtained the highest accuracy at 80:20 data proportions using an epoch of 700, a learning rate of 0.001, a batch size of 25, and some hidden nodes of 100 was 98.41 percent.

show abstract

Section: Introductionmentioning

confidence: 99%

Applying Artificial Neural Network Based on Backpropagation Method for Indonesian Sign Language Recognition

Irvanizam,

Horatius,

Sofyan

2023

IJCDS

View full text Add to dashboard Cite

show abstract

Perturbation-Aided Fiber Nonlinearity Compensation in Subcarrier-Multiplexing Systems

Tang

Liu

et al. 2023

J. Lightwave Technol.

View full text Add to dashboard Cite

Cognitive learning enabled agile optical network

Cheng,

Chen,

et al. 2024

J. Opt. Commun. Netw.

View full text Add to dashboard Cite

Nonlinear equalization (NLE) is essential for guaranteeing the performance of an optical network (ON). Effective NLE implementation relies on key parameters of the transmission link, including the modulation format (MF) and the launch power. As ONs become more agile, the parameters of fiber optical transmission need to be adaptive and relevant to the routing condition. Therefore, successful NLE implementation relies on the realization of transmission awareness (TA). Although machine learning-enabled optical performance monitoring (OPM) has been extensively investigated in the past few years, current NLE algorithms cannot autonomously perceive transmission parameters. Furthermore, current TA implementation still needs human intervention to guide the NLE. In addition, existing ML-based OPM and NLE cannot be trained autonomously, leading to the incapability of environmental change and mislabeling. Here, we propose cognitive learning (CL) for TA-guided NLE in agile ONs. We perform an experiment involving 32 Gbaud polarization-division-multiplexed (PDM)-quadrature phase shift keying (QPSK)/16-quadrature amplitude modulation (QAM) transmission over 1500 km of standard single-mode fiber (SSMF) with a variable launch power from 0 to 3 dBm. When a deep neural network (DNN) with amplitude histograms (AHs) as inputs and one step per span-learned digital back-propagation (1stps-LDBP) are developed, the CL simultaneously enables both TA and NLE, with the capability of self-learning, mislabeling resistance, and dynamic adaptation. The proof-of-concept experimental results indicate that both the accuracy of TA and the Q-factor of PDM-16QAM can be improved by 34.8% and 0.84 dB, respectively, when the launch power is 3 dBm. Moreover, the accuracy of TA is enhanced by 35.3%, even when the used data has 30% mislabeling. Therefore, the CL framework can be customized to satisfy various NLE implementations, thereby supporting the adaptive transmission of agile ONs.

show abstract

Reducing Training Time of Deep Learning Based Digital Backpropagation by Stacking

Cited by 4 publications

References 14 publications

Applying Artificial Neural Network Based on Backpropagation Method for Indonesian Sign Language Recognition

Applying Artificial Neural Network Based on Backpropagation Method for Indonesian Sign Language Recognition

Perturbation-Aided Fiber Nonlinearity Compensation in Subcarrier-Multiplexing Systems

Cognitive learning enabled agile optical network

Contact Info

Product

Resources

About