Information Theoretic Metric Learning (ITML)

An information-theory based distance metric learning algorithm. Given an initial metric, it learns the nearest metric that satisfies some similarity and dissimilarity constraints. The closeness between the metrics is measured using the Kullback-Leibler divergence between the corresponding gaussians.

Watch the full ITML documentation here.

References

Jason V Davis et al. “Information-theoretic metric learning”. In: Proceedings of the 24th international conference on Machine learning. ACM. 2007, pages 209-216.