Mathias Drton: Parameter Identification in Linear Causal Models with Latent Variables
Tid: Ti 2021-03-02 kl 11.15
Plats: Zoom, meeting ID: 625 8662 8413
Medverkande: Mathias Drton, Technical University of Munich
Abstract
Linear causal models relate random variables of interest via a linear equation system that features stochastic noise. These models, also known as structural equation models, are naturally represented by directed graphs whose edges indicate non-zero coefficients in the linear equations. In this talk I will report on progress on combinatorial conditions for parameter identifiability in models with latent (i.e., unobserved) variables. Identifiability holds if the coefficients associated with the edges of the graph can be uniquely recovered from the covariance matrix they define.