Latent Structure
Latent structure research focuses on uncovering hidden, underlying patterns and relationships within complex data, aiming to improve model interpretability, efficiency, and generalization. Current research emphasizes developing methods to learn and leverage these latent structures using diverse approaches, including graph neural networks, diffusion models, and variational autoencoders, often applied to high-dimensional data in various domains like natural language processing, image generation, and biomedical analysis. This work has significant implications for advancing machine learning capabilities, enabling more robust and explainable AI systems, and facilitating deeper understanding of complex phenomena across scientific disciplines.
Papers
November 28, 2022
November 14, 2022
October 22, 2022
October 11, 2022
October 9, 2022
October 5, 2022
August 16, 2022
July 5, 2022
June 13, 2022
March 6, 2022
January 26, 2022
January 3, 2022
December 14, 2021