Learning a Mahalanobis Metric from Equivalence Constraints

Bar-HillelAharon, HertzTomer, ShentalNoam, WeinshallDaphna · Journal of Machine Learning Research · 2005

Many learning algorithms use a metric defined over the input space as a principal tool, and their performance critically depends on the quality of this metric. We address the problem of learning me...

Read the paper · More papers on PaperTik