Maxent Grammar Tool
A program hosted by Bruce Hayes, Dept. of Linguistics, UCLA
What's it for?
This software computes weights for constraint-based maxent grammars. It is meant to be a useful tool for linguists.
The key idea is that you make up a grammar consisting of constraints, then train it, using the program to match a corpus of data. The goal might be to model real-life language learning, or it might simply to make a grammar that's more accurate than one you could produce by hand.
To accomplish this, you need a mathematical expression of the grammar that: (a) lets it make quantitative predictions; (b) reliably (provably) yields the optimum grammar compatible with the data. Maxent is, as far as I know, currently the only mathematical grammar framework for constraint-based theories that satisfy these two criteria.
What does it do?
Using an input file, you feed the program the following
The program computes, and writes to an output file:
Where to learn about maxent
Goldwater, Sharon and Mark Johnson (2003) Learning OT constraint rankings using a Maximum Entropy model. Proceedings of the Workshop on Variation within Optimality Theory, Stockholm University, 2003. Download.
An explanation for the layman on how the weights get found:
Hayes, Bruce and Colin Wilson (2008) "A maximum entropy model of phonotactics and phonotactic learning," Linguistic Inquiry 39: 379-440. Download
Further explanation, with a nice application of the method in linguistics:
Wilson, Colin (2006) Learning phonology with substantive bias: an experimental and computational study of velar palatalization. Cognitive Science 30.5:945-982. Download
Credits
Maxent Grammar Tool originated in software prepared by Colin Wilson for purposes of writing Wilson (2006), cited above. The interface and user-friendliness improvements were carried out by Ben George under a grant from the UCLA Academic Senate Council on Research to Bruce Hayes.