Mathematics

Reports Outline Machine Learning Research from Technion-Israel Institute of Technology

  2012 MAR 27 - (VerticalNews.com) -- According to the authors of recent research from Haifa, Israel, "We derive generalization bounds for learning algorithms based on their robustness: the property that if a testing sample is 'similar' to a training sample, then the testing error is close to the training error. This provides a novel approach, different from complexity or stability arguments, to study generalization of learning algorithms."

  "One advantage of the robustness approach, compared to previous methods, is the geometric intuition it conveys. Consequently, robustness-based analysis is easy to extend to learning in non-standard setups such as Markovian samples or quantile loss," wrote H. Xu and colleagues, Technion-Israel Institute of Technology.

  The researchers concluded: "We further show that a weak notion of robustness is both sufficient and necessary for generalizability, which implies that robustness is a fundamental property that is required for learning algorithms to work."

  Xu and colleagues published their study in Machine Learning (Robustness and generalization. Machine Learning, 2012;86(3):391-423).

  For additional information, contact H. Xu, Technion Israel Inst Technol, Dept. of Elect Engn, IL-32000 Haifa, Israel.

  Publisher contact information for the journal Machine Learning is: Springer, Van Godewijckstraat 30, 3311 Gz Dordrecht, Netherlands.

  Keywords: City:Haifa, Country:Israel, Region:Asia, Cyborgs, Emerging Technologies

  This article was prepared by VerticalNews Mathematics editors from staff and other reports. Copyright 2012, VerticalNews Mathematics via VerticalNews.com.