Graded Grammaticality in Prediction Fractal Machines

Abstract

We introduce a novel method of constructing language models, which avoids some of the problems associated with recurrent neu(cid:173) ral networks. The method of creating a Prediction Fractal Machine (PFM) [1] is briefly described and some experiments are presented which demonstrate the suitability of PFMs for language modeling. PFMs distinguish reliably between minimal pairs, and their be(cid:173) havior is consistent with the hypothesis [4] that wellformedness is 'graded' not absolute. A discussion of their potential to offer fresh insights into language acquisition and processing follows.

Cite

Text

Parfitt et al. "Graded Grammaticality in Prediction Fractal Machines." Neural Information Processing Systems, 1999.

Markdown

[Parfitt et al. "Graded Grammaticality in Prediction Fractal Machines." Neural Information Processing Systems, 1999.](https://mlanthology.org/neurips/1999/parfitt1999neurips-graded/)

BibTeX

@inproceedings{parfitt1999neurips-graded,
  title     = {{Graded Grammaticality in Prediction Fractal Machines}},
  author    = {Parfitt, Shan and Tiño, Peter and Dorffner, Georg},
  booktitle = {Neural Information Processing Systems},
  year      = {1999},
  pages     = {52-58},
  url       = {https://mlanthology.org/neurips/1999/parfitt1999neurips-graded/}
}