search

roylanceMichael / cs_6340_nlp_coreference_utah

uu nlp coreference
2 stars 1 forks source link

Spark Stanford API, Wordnet, and resources #4

Closed illegalnumbers closed 11 years ago

illegalnumbers commented 11 years ago

Find out more info on the stanford parser classes as well as the coreference resources Prof Riloff posted and the Wordnet synonym sets. Do this to tweak the algorithm to do better than just 'guessing' coref antecedants.

illegalnumbers commented 11 years ago

I found one of the algorithms in the resources Ellen listed to be really cool and not dependent on a large corpus or a lot of machine learning, I will talk with you about it and get your feelings on it when we meet on Saturday. I recommend you take a look at the papers Ellen listed if you have time, some of them are pretty cool and all are pretty applicable to the project. I will continue looking at the Stanford API, but since you can already parse out the NP's I will not dedicate much time to this. I have looked at Wordnet and it's a cool resource but I guess it really depends on the approach of our tweaking of the project to get everything running if we want to use it or not.