This version improves parsing speed using the hash kernel (see [4]) and by optimizing the code. We also improved the unlabeled attachment score (UAS) slightly and labeled attachment score (LAS) significantly.
#### [Quick Start](https://github.com/taolei87/RBGParser/wiki/Quick-Start)
#### [FAQs](https://github.com/taolei87/RBGParser/wiki/FAQs)
======
This project is developed at Natural Language Processing group in MIT. It contains a Java implementation of a syntactic dependency parser with tensor decomposition and greedy decoding, described in [1,2,3].
This project is implemented by Tao Lei (taolei [at] csail.mit.edu) and Yuan Zhang (yuanzh [at] csail.mit.edu).
=========
[1] Tao Lei, Yu Xin, Yuan Zhang, Regina Barzilay and Tommi Jaakkola. Low-Rank Tensors for Scoring Dependency Structures. ACL 2014. PDF
[2] Yuan Zhang, Tao Lei, Regina Barzilay, Tommi Jaakkola and Amir Globerson. Steps to Excellence: Simple Inference with Refined Scoring of Dependency Trees. ACL 2014. PDF
[3] Yuan Zhang*, Tao Lei*, Regina Barzilay and Tommi Jaakkola. Greed is Good if Randomized: New Inference for Dependency Parsing. EMNLP 2014. PDF
[4] Bernd Bohnet. Very High Accuracy and Fast Dependency Parsing is not a Contradiction. The 23rd International Conference on Computational Linguistics. COLING 2010. PDF