Learning with Side Information (LSI)

Also known as MMC (Mahalanobis Metric for Clustering), this distance metric learning learns a metric that globally minimizes the distance between similar points, with the constraint that dissimilar points must be far enough. This algorithm can be used for supervised learning, but is also valid for clustering with side information.

Watch the full LSI documentation here.

References

Eric P Xing et al. “Distance metric learning with application to clustering with side-information”. In: Advances in neural information processing systems. 2003, pages 521-528.