GNN Dependency Parser

The code of "Graph-based Dependency Parsing with Graph Neural Networks".

Requirements

• python: 3.6.0
• dynet: 2.0.0
• antu: 0.0.5

Example log

An example of experiment log.

Training

$ cd src
$ python train.py --config_file ../configs/default.cfg --name ACL19(your experiment name) --gpu 0(your gpu id)

Before triggering the subcommands, please make sure that the data files must be in CoNLL-U format. Here is an example.

$ cat data/dev.debug 
1   Influential _   JJ  JJ  _   2   amod    _   _
2   members _   NNS NNS _   10  nsubj   _   _
3   of  _   IN  IN  _   2   prep    _   _
4   the _   DT  DT  _   6   det _   _
5   House   _   NNP NNP _   6   nn  _   _
6   Ways    _   NNP NNP _   3   pobj    _   _
7   and _   CC  CC  _   6   cc  _   _
8   Means   _   NNP NNP _   9   nn  _   _
9   Committee   _   NNP NNP _   6   conj    _   _
10  introduced  _   VBD VBD _   0   root    _   _
11  legislation _   NN  NN  _   10  dobj    _   _
12  that    _   WDT WDT _   14  nsubj   _   _
13  would   _   MD  MD  _   14  aux _   _
14  restrict    _   VB  VB  _   11  rcmod   _   _
15  how _   WRB WRB _   22  advmod  _   _
16  the _   DT  DT  _   20  det _   _
17  new _   JJ  JJ  _   20  amod    _   _
18  savings-and-loan    _   NN  JJ  _   20  nn  _   _
19  bailout _   NN  NN  _   20  nn  _   _
20  agency  _   NN  NN  _   22  nsubj   _   _
21  can _   MD  MD  _   22  aux _   _
22  raise   _   VB  VB  _   14  ccomp   _   _
23  capital _   NN  NN  _   22  dobj    _   _
24  ,   _   ,   ,   _   14  punct   _   _
25  creating    _   VBG VBG _   14  xcomp   _   _
26  another _   DT  DT  _   28  det _   _
27  potential   _   JJ  JJ  _   28  amod    _   _
28  obstacle    _   NN  NN  _   25  dobj    _   _
29  to  _   TO  TO  _   28  prep    _   _
30  the _   DT  DT  _   31  det _   _
31  government  _   NN  NN  _   33  poss    _   _
32  's  _   POS POS _   31  possessive  _   _
33  sale    _   NN  NN  _   29  pobj    _   _
34  of  _   IN  IN  _   33  prep    _   _
35  sick    _   JJ  JJ  _   36  amod    _   _
36  thrifts _   NNS NNS _   34  pobj    _   _
37  .   _   .   .   _   10  punct   _   _

Predict

$ cd src
$ python predict.py --config_file ../configs/default.cfg --name PTB-Out(your experiment name) --gpu 0(your gpu id)

Cite

If you find our code is useful, please cite:

@inproceedings{ji-etal-2019-graph,
    title = "Graph-based Dependency Parsing with Graph Neural Networks",
    author = "Ji, Tao  and
      Wu, Yuanbin  and
      Lan, Man",
    booktitle = "Proceedings of the 57th Annual Meeting of the Association for Computational Linguistics",
    month = jul,
    year = "2019",
    address = "Florence, Italy",
    publisher = "Association for Computational Linguistics",
    url = "https://www.aclweb.org/anthology/P19-1237",
    doi = "10.18653/v1/P19-1237",
    pages = "2475--2485",
}