Probabilistic Graphical Model
Probabilistic graphical models (PGMs) are statistical frameworks representing complex systems through graphs where nodes denote variables and edges encode probabilistic dependencies. Current research emphasizes efficient inference algorithms, particularly for large-scale models and those incorporating latent variables, with a focus on improving scalability and accuracy through techniques like lifted inference, variational methods, and the integration of deep learning architectures. PGMs find broad application in diverse fields, from causal inference and knowledge graph analysis to semantic communication and anomaly detection, offering powerful tools for modeling uncertainty and extracting insights from data.
Papers
PGMax: Factor Graphs for Discrete Probabilistic Graphical Models and Loopy Belief Propagation in JAX
Guangyao Zhou, Antoine Dedieu, Nishanth Kumar, Wolfgang Lehrach, Miguel Lázaro-Gredilla, Shrinu Kushagra, Dileep George
Hierarchical Dependency Constrained Tree Augmented Naive Bayes Classifiers for Hierarchical Feature Spaces
Cen Wan, Alex A. Freitas