Corpus ID: 1585615

Learning Low Density Separators

  title={Learning Low Density Separators},
  author={Shai Ben-David and Tyler Lu and D. P{\'a}l and Miroslava Sot{\'a}kov{\'a}},
  • Shai Ben-David, Tyler Lu, +1 author Miroslava Sotáková
  • Published 2009
  • Computer Science, Mathematics
  • ArXiv
  • We define a novel, basic, unsupervised learning problem – learning hyperplane passing through the origin with the lowest probability density. Namely, given a random sample generated by some unknown probability distribution, the task is to find a hyperplane passing through the origin with smallest integral of the probability density on the hyperplane. This task is relevant to several problems in machine learning, such as semisupervised learning and clustering stability. We investigate the… CONTINUE READING
    19 Citations
    Minimum Density Hyperplanes
    • 17
    • Highly Influenced
    • PDF
    Minimum density hyperplanes in the feature space
    • 1
    • PDF
    Minimum Density Hyperplane: An Unsupervised and Semi-Supervised Classifier
    • 4
    • Highly Influenced
    • PDF
    Minimum spectral connectivity projection pursuit
    • 4
    • PDF
    Minimum Conditional Entropy Clustering: A Discriminative Framework for Clustering
    • 7
    • Highly Influenced
    • PDF
    Fundamental Limitations of Semi-Supervised Learning
    • 18
    Projection methods for clustering and semi-supervised classification
    • PDF
    Combining Low-Density Separators with CNNs
    • PDF


    The Computational Complexity of Densest Region Detection
    • 25
    • PDF
    On nonparametric estimation of density level sets
    • 197
    Adaptive Hausdorff Estimation of Density Level Sets
    • 66
    • PDF
    Neural Network Learning - Theoretical Foundations
    • 1,392
    Relating Clustering Stability to Properties of Cluster Boundaries
    • 48
    • PDF
    Randomized Algorithms
    • 1,759
    Combinatorial methods in density estimation
    • 694
    • PDF