This paper proposes a machine learning approach to estimate physical forward default intensities. Default probabilities are computed using artificial neural networks to estimate the intensities of the inhomogeneous Poisson processes governing default process. The major contribution to previous...