Architecture Knowledge Distillation for Evolutionary Generative Adversarial Network
International Journal of Neural Systems · 2025-04 · 35 (04)
DOI: 10.1142/s0129065725500133 ↗Kurz gefasst
Fachartikel in der Zeitschrift International Journal of Neural Systems, online erschienen am 19.02.2025. Verfasst von Yu Xue und 2 weiteren Autor:innen. Beteiligte Einrichtung: School of Computer and Software, Nanjing University of Information Science and Technology, Nanjing 210044, P. R. China. Hauptthemen laut MEDLINE: Neural Networks, Computer, Image Processing, Computer-Assisted. Die Arbeit wurde bisher 22-mal zitiert (Crossref) und nennt 75 Literaturangaben. Gefördert von: National Natural Science Foundation of China, Natural Science Foundation of Jiangsu Province, Natural Science Foundation of the Jiangsu Higher Education Institutions of China.
Abstract (Originalsprache)
Generative Adversarial Networks (GANs) are effective for image generation, but their unstable training limits broader applications. Additionally, neural architecture search (NAS) for GANs with one-shot models often leads to insufficient subnet training, where subnets inherit weights from a supernet without proper optimization, further degrading performance. To address both issues, we propose Architecture Knowledge Distillation for Evolutionary GAN (AKD-EGAN). AKD-EGAN operates in two stages. First, architecture knowledge distillation (AKD) is used during supernet training to efficiently optimize subnetworks and accelerate learning. Second, a multi-objective evolutionary algorithm (MOEA) searches for optimal subnet architectures, ensuring efficiency by considering multiple performance metrics. This approach, combined with a strategy for architecture inheritance, enhances GAN stability and image quality. Experiments show that AKD-EGAN surpasses state-of-the-art methods, achieving a Fréchet Inception Distance (FID) of 7.91 and an Inception Score (IS) of 8.97 on CIFAR-10, along with competitive results on STL-10 (FID: 20.32, IS: 10.06). Code and models will be available at https://github.com/njit-ly/AKD-EGAN .
Schlagwörter
MeSH-Schlagwörter (MEDLINE)
Stichwörter der Autor:innen
Autor:innen und Institutionen
- School of Computer and Software, Nanjing University of Information Science and Technology, Nanjing 210044, P. R. China.
- NICE Research Group, School of Computer Science and Electronic Engineering, University of Surrey, Guildford, GU2 7XS, UK.
ORCID-Profile: Yu Xue, Yan Lin, Ferrante Neri
Förderung
- National Natural Science Foundation of China
- Natural Science Foundation of Jiangsu Province
- Natural Science Foundation of the Jiangsu Higher Education Institutions of China
Zitieren
APA
Xue, Y., Lin, Y., & Neri, F. (2025). Architecture Knowledge Distillation for Evolutionary Generative Adversarial Network. International Journal of Neural Systems, 35(04). https://doi.org/10.1142/s0129065725500133
Vancouver
Xue Y, Lin Y, Neri F. Architecture Knowledge Distillation for Evolutionary Generative Adversarial Network. International Journal of Neural Systems. 2025;35(04). doi:10.1142/s0129065725500133
BibTeX
@article{Xue2025,
author = {Xue, Yu and Lin, Yan and Neri, Ferrante},
title = {Architecture Knowledge Distillation for Evolutionary Generative Adversarial Network},
journal = {International Journal of Neural Systems},
year = {2025},
volume = {35},
number = {04},
doi = {10.1142/s0129065725500133}
}
EUVIMED