Improving Maximum Likelihood Estimation Using Marginalization and Black-Box Variational Inference
Based upon Black Box Variational Inference, a new set of classification algorithms has recently emerged. The goals of this set of algorithms are twofold: 1) increasing generalization power; 2) decreasing computational and implementation complexity. To this end, we assume a set of latent variables during the generation of data points. We subsequently marginalize the conventional classification likelihood objective function w.r.t this set of latent variables and then apply black-box variational inference to estimate the marginalized likelihood. We evaluate the performance of the proposed method by comparing the results obtained from the application of our method to real-world datasets with those obtained using several classification algorithms. The experimental results prove that our proposed method is competitive.