Abstract - Machine Learning || Yoav Goldberg

Phd Student at Ben Gurion University



Challenges of Syntactic Processing of Hebrew, and advances in Dependency Parsing

I will describe my experience with designing a syntactic parser for Hebrew, a language with rich morphology and a small treebank. After describing some of the characteristics that make automatic syntactic processing of Hebrew challanging, I will present some solutions to these challenges. The work on Hebrew brought about solutions that work well also for English -- I will touch upon them as well. I will present a constituency parsing system that performs joint morphological segmentation and syntactic parsing, and a greedy dependency parsing system which is fast, accurate, and can accommodate a range of complex features. Time permitting, I will touch upon the issue of confidence estimation for the output of dependency parsers.