On-line learning

From Scholarpedia

This article has not been peer-reviewed or accepted for publication yet; It may be unfinished, contain inaccuracies, or unapproved changes.

Author: Dr. Leon Bottou, NEC Laboratories America, Princeton, NJ

While this article is empty, see On-line learning on Amazon.

Dr. Leon Bottou accepted the invitation on 15 May 2007.


This article will briefly cover: Learning algorithms that work by successively processing fresh examples.

  • origins and usage of on-line learning algorithms.
  • perceptron algorithm, regret bounds.
  • stochastic gradient algorithms, stochastic approximations.
  • optimal on-line learning, second order efficiency.
  • conclusion

Potential advanced topics:

  • duality and on-line learning?
  • on-line learning and independence?
Invited by: Dr. Eugene M. Izhikevich, Editor-in-Chief of Scholarpedia, the peer-reviewed open-access encyclopedia
Retrieved from "http://www.scholarpedia.org/article/On-line_learning"
For authors