Unsupervised Learning with Imbalanced Data via Structure Consolidation Latent Variable Model

Fariba Yousefi, Zhenwen Dai, Carl Henrik Ek, Neil Lawrence

Abstract: Unsupervised learning on imbalanced data is challenging because, when given imbalanced data, current model is often dominated by the major category and ignores the categories with small amount of data. We develop a latent variable model that can cope with imbalanced data by dividing the latent space into a shared space and a private space. Based on Gaussian Process Latent Variable Models, we propose a new kernel formulation that enables the separation of latent space and derive an efficient variational inference method. The performance of our model is demonstrated with an imbalanced medical image dataset.

Conflicts: sheffield.ac.uk, xrce.xerox.com, fias.uni-frankfurt.de