Category Ranking
98%
Total Visits
921
Avg Visit Duration
2 minutes
Citations
20
Article Abstract
We propose an online knowledge distillation method using a teacher-student model for nonlinear equalization in high-symbol-rate dense wavelength division multiplexing (DWDM) system spanning the S+C+L bands. The system features 195 WDM channels with 100 GHz spacing, transmitting 98 GBaud polarization division multiplexing (PDM) probability constellation shaping (PCS) 256-quadrature amplitude modulation (QAM) signals over 150 km of G.654.E standard single mode fiber (SSMF). Teacher models trained on five representative C-band channels are distilled to one or two lightweight student models on each of the C, S, L bands to achieve the whole triple-band equalization. This approach improves the system's net data-rate from 201.5 Tb/s to 231.6 Tb/s, and reduces computational complexity by up to 97.1% in C band and 80% in S and L band. The results highlight that the student model can achieve comparable performance to teacher models with significantly reduced complexity.
Download full-text PDF |
Source |
|---|---|
| http://dx.doi.org/10.1364/OE.559025 | DOI Listing |
Publication Analysis
Top Keywords
Similar Publications
DRFNet: Enhancing Identity Discriminability and Feature Robustness for Cross-Session VEP-Based EEG Biometrics.
IEEE J Biomed Health Inform
September 2025
Biometric recognition using visually evoked potentials (VEPs), a type of neural response to visual stimuli recorded via electroencephalography (EEG), has shown great promise. However, the non-stationary nature of EEG signals poses a major challenge in cross-session scenarios, where data collected on different days often leads to performance degradation. To address this, we propose the Discriminative Robust Feature Network (DRFNet) to enhance the robustness and inter-subject discriminability of identity representations across sessions.
View Article and Find Full Text PDFAssociation of Prognostic Nutritional Index and Mortality in Older Adults Undergoing Hip Fracture Surgery: A Retrospective Observational Study at a Single Large Center.
Medicina (Kaunas)
July 2025
Department of Anesthesiology and Pain Medicine, Asan Medical Center, University of Ulsan College of Medicine, 88 Olympic-ro 43-gil, Songpa-gu, Seoul 05505, Republic of Korea.
: Patients with hip fractures have a high mortality rate, highlighting the need for a reliable prognostic tool. Although the prognostic nutritional index (PNI) is a well-established predictor in patients with cancer, its utility has not been thoroughly investigated in patients with hip fractures. Therefore, this study aims to evaluate the association between PNI and mortality in patients undergoing hip fracture surgery.
View Article and Find Full Text PDFFrom information exposure to legal consciousness: The chain mediating role of political trust and social equality in Chinese college students.
Acta Psychol (Amst)
September 2025
School of Electrical and Electronic Information, Xihua University, Chengdu, Sichuan province 610039, China.
The cultivation of college students' legal consciousness in the digital media environment is crucial for advancing social responsibility, moral awareness, and rule-of-law societies. However, prior research often assumes a linear positive relationship between information exposure and legal consciousness, overlooking potential negative effects in which algorithmic biases and emotional framing in digital media may weaken legal consciousness. Existing studies rarely explore the chain mediation of political trust and social equality perceptions, particularly lacking empirical tests on Chinese college students.
View Article and Find Full Text PDFChaotic chimp-mountain gazelle optimized FOPID control for frequency regulation in islanded airport microgrids with heterogeneous energy systems.
Sci Rep
August 2025
Department of Electrical Power and Control Engineering, School of Electrical Engineering and Computing, Adama Science and Technology University, Adama, Ethiopia.
Ensuring stable frequency regulation in islanded airport microgrids is a challenging task owing to the intermittency of renewable energy sources, unpredictable load variations, and nonlinear dynamics. Conventional optimization techniques often struggle with premature convergence and sub-optimal controller tuning, leading to a poor transient response and inadequate frequency stabilization. These challenges necessitate an advanced optimization strategy that can efficiently handle dynamic airport environments while ensuring enhanced frequency stability.
View Article and Find Full Text PDFCrop yield and water productivity modeling using nonlinear growth functions.
Sci Rep
August 2025
Department of Irrigation and Reclamation Engineering, College of Agriculture and Natural Resources, University of Tehran, Karaj, Iran.
Growth curve modeling plays a crucial role in precision agriculture by enabling rapid analysis of plant growth dynamics. Understanding the complex mechanisms of crop growth is essential for optimizing agricultural productivity. In this study, nonlinear Logistic and Gompertz models were employed to predict biological yield and water productivity of silage maize in arid and semi-arid regions, using growing degree days (GDD) as a key predictor.
View Article and Find Full Text PDF