ML Anthology

A Comparison of Learning Techniques in Second Language Learning

Steven L. Lytinen, Carol E. Moon
ICML 1990 pp. 377-383
doi:10.1016/B978-1-55860-141-3.50048-1 /icml/1990/lytinen1990icml-comparison/

Abstract

We present a machine learning program, called ANT, which learns the grammar of a second language. Input to the system is similar to what is found in a typical introductory foreign language text; that is, a mixture of instructions describing grammar rules, and examples illustrating these rules. We compare ANT's learning to two alternatives: learning from only instructions, and learning from only examples. We discuss why, from a. functional or processing standpoint, learning from a mixed input is more effective than either of the alternatives. We also present an empirical comparison of our algorithm's performance on input containing both instructions and examples vs. performance of the system when given instructions only or examples only. The results of the comparison support our hypotheses as to the utility of mixed input.

ICML Semantic Scholar

Cite

Text

Lytinen and Moon. "A Comparison of Learning Techniques in Second Language Learning." International Conference on Machine Learning, 1990. doi:10.1016/B978-1-55860-141-3.50048-1

Markdown

[Lytinen and Moon. "A Comparison of Learning Techniques in Second Language Learning." International Conference on Machine Learning, 1990.](https://mlanthology.org/icml/1990/lytinen1990icml-comparison/) doi:10.1016/B978-1-55860-141-3.50048-1

BibTeX

@inproceedings{lytinen1990icml-comparison,
  title     = {{A Comparison of Learning Techniques in Second Language Learning}},
  author    = {Lytinen, Steven L. and Moon, Carol E.},
  booktitle = {International Conference on Machine Learning},
  year      = {1990},
  pages     = {377-383},
  doi       = {10.1016/B978-1-55860-141-3.50048-1},
  url       = {https://mlanthology.org/icml/1990/lytinen1990icml-comparison/}
}