Litcius/Paper detail

Architecture Knowledge Distillation for Evolutionary Generative Adversarial Network

Yu Xue, Yan Lin, Ferrante Neri

2024International Journal of Neural Systems17 citationsDOI

Abstract

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.

Topics & Concepts

SubnetComputer scienceDistillationArchitectureArtificial intelligenceArtificial neural networkMachine learningNetwork architectureComputer securityVisual artsComputer networkArtOrganic chemistryChemistryGenerative Adversarial Networks and Image SynthesisAdvanced Image Processing TechniquesAdvanced Vision and Imaging
Architecture Knowledge Distillation for Evolutionary Generative Adversarial Network | Litcius