Generate more than one child in your co-evolutionary semi-supervised learning GAN

2025-04-29

AI Summary

Existing co-evolutionary SSL-GAN approaches suffer from low evolutionary efficiency and insufficient diversity due to single-offspring generation and cellular population structures. To address these limitations, we propose a novel cooperative evolutionary framework that replaces spatially constrained populations with a panmictic (fully mixed) structure, enables parallel multi-offspring generation per generation, and incorporates an elite preservation strategy for efficient selection and stable convergence. This work is the first to jointly integrate panmixia, multi-offspring mutation, and elite replacement into SSL-GAN co-evolution, thereby breaking the conventional paradigm of single-generation, sequential update. Extensive experiments on CIFAR-10, SVHN, and STL-10 demonstrate significant improvements in classification accuracy under semi-supervised settings, alongside enhanced training stability. These results validate the effectiveness and advancement of our framework for semi-supervised generative adversarial learning.

Technology Category

• Machine Learning: Semi-Supervised Learning
• Computer Vision: Generative Adversarial Networks (GANs) for Vision
• Search and Optimization: Evolutionary Computation

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Web Mining and Content Analysis: Models for Web evolution
• Economics, Online Markets and Human Computation: Economic ramifications for generative AI infrastructure and applications

Abstract

Generative Adversarial Networks (GANs) are very useful methods to address semi-supervised learning (SSL) datasets, thanks to their ability to generate samples similar to real data. This approach, called SSL-GAN has attracted many researchers in the last decade. Evolutionary algorithms have been used to guide the evolution and training of SSL-GANs with great success. In particular, several co-evolutionary approaches have been applied where the two networks of a GAN (the generator and the discriminator) are evolved in separate populations. The co-evolutionary approaches published to date assume some spatial structure of the populations, based on the ideas of cellular evolutionary algorithms. They also create one single individual per generation and follow a generational replacement strategy in the evolution. In this paper, we re-consider those algorithmic design decisions and propose a new co-evolutionary approach, called Co-evolutionary Elitist SSL-GAN (CE-SSLGAN), with panmictic population, elitist replacement, and more than one individual in the offspring. We evaluate the performance of our proposed method using three standard benchmark datasets. The results show that creating more than one offspring per population and using elitism improves the results in comparison with a classical SSL-GAN.

Problem

• Enhancing co-evolutionary SSL-GANs with multi-offspring generation
• Improving GAN performance via panmictic populations and elitism
• Optimizing semi-supervised learning using advanced evolutionary strategies

Innovation

• Co-evolutionary GAN with panmictic population
• Elitist replacement strategy in evolution
• Multiple offspring individuals per generation