Distribution Learning

Distribution learning focuses on estimating probability density functions from data samples, a crucial task with applications across diverse fields. Current research emphasizes developing robust and efficient algorithms for various distribution types, including Gaussian mixtures and those with complex structures like graphs, often leveraging deep learning architectures such as diffusion models, neural ODEs, and autoencoders. These advancements improve the accuracy and efficiency of tasks ranging from image generation and classification to molecular property prediction and causal inference, impacting both theoretical understanding and practical applications in machine learning and beyond.