Track papers across labs, topics, and trends.
Search papers, labs, and topics across Lattice.
This paper analyzes maximum likelihood variational inference (MLVI) for stochastic block models (SBMs) through the lens of optimal transport (OT), showing that MLVI is equivalent to a semi-relaxed Gromov-Wasserstein (srGW) projection with entropic regularization. They prove that unregularized srGW estimators consistently recover the SBM connectivity matrix and latent cluster assignments asymptotically, but this doesn't translate to reliable model selection in finite samples. To address this, they propose a regularized srGW formulation that promotes sparsity in inferred cluster proportions, enabling simultaneous model parameter recovery and cluster number selection.
Ditching entropic regularization in Gromov-Wasserstein unlocks consistent recovery of SBM connectivity and latent clusters, but you'll need sparsity-promoting regularization for reliable finite-sample model selection.
We study inference in stochastic block models (SBMs) through the lens of optimal transport (OT). We first establish that maximum likelihood variational inference (MLVI) can be interpreted as a semi-relaxed Gromov-Wasserstein (srGW) projection with entropic regularization. While this formulation yields accurate clustering, the entropic regularization prevents transport plans to be sparse, hindering intrinsic model selection. Consequently, we investigate unregularized srGW estimators, and prove that they consistently recover both the SBM connectivity matrix and latent cluster assignments in the asymptotic regime. However, this asymptotic property does not translate into reliable model selection in finite samples, and calls for additional mechanisms to promote sparsity in the inferred cluster proportions. We empirically show that such a regularized formulation yields estimators that simultaneously recover model parameters and select the number of clusters in a single optimization problem, thereby avoiding costly grid search or heuristic model selection procedures.