Learning Low-Density Separators Export

@misc{citeulike:2815247, abstract = {We define a novel, basic, unsupervised learning problem - learning the lowest density homogeneous hyperplane separator of an unknown probability distribution. This task is relevant to several problems in machine learning, such as semi-supervised learning and clustering stability. We investigate the question of existence of a universally consistent algorithm for this problem. We propose two natural learning paradigms and prove that, on input unlabeled random samples generated by any member of a rich family of distributions, they are guaranteed to converge to the optimal separator for that distribution. We complement this result by showing that no learning algorithm for our task can achieve uniform learning rates (that are independent of the data generating distribution).}, archivePrefix = {arXiv}, author = {Ben-David, Shai and Lu, Tyler and Pal, David and Sotakova, Miroslava}, citeulike-article-id = {2815247}, citeulike-linkout-0 = {http://arxiv.org/abs/0805.2891}, citeulike-linkout-1 = {http://arxiv.org/pdf/0805.2891}, eprint = {0805.2891}, keywords = {density\_estimation}, month = {May}, posted-at = {2008-07-21 08:20:59}, priority = {2}, title = {Learning Low-Density Separators}, url = {http://arxiv.org/abs/0805.2891}, year = {2008} }

TY - ELEC ID - citeulike:2815247 L3 - citeulike-article-id:2815247 N2 - We define a novel, basic, unsupervised learning problem - learning the lowest density homogeneous hyperplane separator of an unknown probability distribution. This task is relevant to several problems in machine learning, such as semi-supervised learning and clustering stability. We investigate the question of existence of a universally consistent algorithm for this problem. We propose two natural learning paradigms and prove that, on input unlabeled random samples generated by any member of a rich family of distributions, they are guaranteed to converge to the optimal separator for that distribution. We complement this result by showing that no learning algorithm for our task can achieve uniform learning rates (that are independent of the data generating distribution). TI - Learning Low-Density Separators KW - density_estimation AU - Ben-David, Shai AU - Lu, Tyler AU - Pal, David AU - Sotakova, Miroslava PY - 2008/May/19/ UR - http://arxiv.org/abs/0805.2891 ER -