During my PhD I worked on a new kind of neural network and empirically tested its ability to learn to parse English sentences drawn from a corpus of naturally occurring text.

A Simple Synchrony Network is an extension of regular feed-forward and simple recurrent networks to provide additional representational power. In essence, the output of the network is also given a time by which entities can be identified. In the context of natural language, these entities represent nodes within the parse tree.

Main publications:

  1. P.C.R. Lane and J.B. Henderson, ‘Towards effective parsing with neural networks: Inherent generalisations and bounded resource effects,’ Applied Intelligence, 19:83-100, 2003. Web page.
  2. P.C.R. Lane and J.B. Henderson, ‘Incremental syntactic parsing of natural language corpora with Simple Synchrony Networks,’ IEEE Transactions on Knowledge and Data Engineering, 13:219-231, 2001. Web page.

This work was continued by my supervisor Dr. James Henderson.