Sparse Causal
Sparse causal discovery focuses on identifying cause-and-effect relationships from data where only a limited number of variables directly influence each other, represented as sparse directed acyclic graphs (DAGs). Current research emphasizes developing robust and scalable algorithms, including differentiable methods and those leveraging structural equation models or Hawkes processes, to infer these sparse causal structures from observational and interventional data, even in high-dimensional settings. This work is significant because accurately identifying sparse causal relationships is crucial for understanding complex systems in various fields, improving the accuracy of predictions, and enabling targeted interventions.
Papers
October 2, 2024
March 18, 2024
January 10, 2024
November 17, 2023
May 11, 2023
August 23, 2022