Generative AI in Education: Developing Personalized Learning Experiences with Hyperspherical Variational Self-Attention Autoencoder

With the advancement of artificial intelligence (AI) technology, content generation has sparked a transformative revolution in the field of education. In traditional education, teachers often spend a lot of time preparing lessons, because students have varying levels of proficiency, teachers need to consider whether the teaching content is suitable for everyone. Generative AI provides an effective solution to this problem by automatically generating personalized, high-quality educational content, which not only alleviates the burden on teachers but also enhances students' learning outcomes. This study focuses on a novel deep generative model—a framework based on Variational Autoencoders (VAE) and the self-attention mechanism (Transformer). We propose a model called the Hyperspherical Variational Self-Attention Autoencoder (HVSAE), which aims to generate personalized learning content based on students' learning situations, thereby reducing the burden on teachers and improving learning outcomes. The experimental results indicate that the model can generate high-quality educational resources, providing important support for achieving more personalized and efficient education.