ML Anthology

A Conservation Law for Generalization Performance

Cullen Schaffer
ICML 1994 pp. 259-265
doi:10.1016/B978-1-55860-335-6.50039-8 /icml/1994/schaffer1994icml-conservation/

Abstract

Many aspects of concept learning research can be understood more clearly in light of a basic mathematical result stating, essentially, that positive performance in some learning situations must be offset by an equal degree of negative performance in others. We present a proof of this result and comment on some of its theoretical and practical ramifications.

ICML Semantic Scholar

Cite

Text

Schaffer. "A Conservation Law for Generalization Performance." International Conference on Machine Learning, 1994. doi:10.1016/B978-1-55860-335-6.50039-8

Markdown

[Schaffer. "A Conservation Law for Generalization Performance." International Conference on Machine Learning, 1994.](https://mlanthology.org/icml/1994/schaffer1994icml-conservation/) doi:10.1016/B978-1-55860-335-6.50039-8

BibTeX

@inproceedings{schaffer1994icml-conservation,
  title     = {{A Conservation Law for Generalization Performance}},
  author    = {Schaffer, Cullen},
  booktitle = {International Conference on Machine Learning},
  year      = {1994},
  pages     = {259-265},
  doi       = {10.1016/B978-1-55860-335-6.50039-8},
  url       = {https://mlanthology.org/icml/1994/schaffer1994icml-conservation/}
}