eduzhai > Applied Sciences > Computer Science >

Towards Incremental Parsing of Natural Language using Recursive Neural Networks

  • paiqiu
  • (0) Download
  • 20210804
  • Save

... pages left unread,continue reading

Document pages: 30 pages

Abstract: In this paper we develop novel algorithmic ideas for building a natural languageparser grounded upon the hypothesis of incrementality. Although widely acceptedand experimentally supported under a cognitive perspective as a model of the humanparser, the incrementality assumption has never been exploited for building automaticparsers of unconstrained real texts. The essentials of the hypothesis are that words areprocessed in a left-to-right fashion, and the syntactic structure is kept totally connectedat each step.Our proposal relies on a machine learning technique for predicting the correctness ofpartial syntactic structures that are built during the parsing process. A recursive neuralnetwork architecture is employed for computing predictions after a training phase onexamples drawn from a corpus of parsed sentences, the Penn Treebank. Our resultsindicate the viability of the approach andlay out the premises for a novel generation ofalgorithms for natural language processing which more closely model human parsing.These algorithms may prove very useful in the development of eÆcient parsers.

Please select stars to rate!

         

0 comments Sign in to leave a comment.

    Data loading, please wait...
×