Masahiro Kohjima

We propose a probabilistic model using Gaussian processes for learning regression functions from data without correspondence.