Agnostic Pointwise-Competitive Selective Classification

@article{Wiener2015AgnosticPS,
  title={Agnostic Pointwise-Competitive Selective Classification},
  author={Yair Wiener and Ran El-Yaniv},
  journal={J. Artif. Intell. Res.},
  year={2015},
  volume={52},
  pages={171-201}
}
A pointwise competitive classifier from class F is required to classify identically to the best classifier in hindsight from F. For noisy, agnostic settings we present a strategy for learning pointwise-competitive classifiers from a finite training sample provided that the classifier can abstain from prediction at a certain region of its choice. For some interesting hypothesis classes and families of distributions, the measure of this rejected region is shown to be diminishing at rate β1 ċ O… CONTINUE READING

References

Publications referenced by this paper.
SHOWING 1-10 OF 71 REFERENCES

On the Foundations of Noise-free Selective Classification

  • J. Mach. Learn. Res.
  • 2010
VIEW 21 EXCERPTS
HIGHLY INFLUENTIAL

Theoretical foundations of active learning

VIEW 6 EXCERPTS
HIGHLY INFLUENTIAL

Classification with a Reject Option using a Hinge Loss

  • J. Mach. Learn. Res.
  • 2008
VIEW 13 EXCERPTS
HIGHLY INFLUENTIAL

Introduction to Statistical Learning Theory

  • Advanced Lectures on Machine Learning
  • 2003
VIEW 7 EXCERPTS
HIGHLY INFLUENTIAL

Theoretical Foundations of Selective Prediction

Y. Wiener
  • Ph.D. thesis, Technion — Israel Institute of Technology.
  • 2013
VIEW 5 EXCERPTS
HIGHLY INFLUENTIAL

Active Learning via Perfect Selective Classification

  • J. Mach. Learn. Res.
  • 2012
VIEW 6 EXCERPTS
HIGHLY INFLUENTIAL

Similar Papers

Loading similar papers…