Xueyan Jiang

Papers

Total Citations

H-Index

About

Xueyan Jiang is a leading researcher in the field of deep generative modeling, with a particular focus on developing rigorous evaluation frameworks for complex neural architectures. Their seminal 2017 work, "Metrics for Deep Generative Models," has garnered 36 citations and established foundational principles for assessing how well models like variational autoencoders (VAEs) and generative adversarial networks (GANs) approximate target distributions. Jiang’s research critically examines the manifold hypothesis, exploring how neural samplers transform latent spaces into high-dimensional data representations. By proposing systematic metrics for evaluating generative fidelity and diversity, they have helped standardize performance comparisons across competing approaches, enabling more reliable benchmarking in the field. Their contributions address a fundamental challenge in unsupervised learning: quantifying how effectively a model captures the true underlying structure of complex datasets. Jiang’s work continues to influence both theoretical understanding and practical applications of generative AI, providing essential tools for researchers developing more robust and interpretable deep learning systems. Their insights into latent space geometry and distributional alignment remain highly relevant as the field advances toward more sophisticated generative architectures.