HRL-RE

Codes for the paper "A Hierarchical Framework for Relation Extraction with Reinforcement Learning", and you can find our paper at arxiv

Cite this paper :

@inproceedings{takanobu2019hierarchical,
  title={A Hierarchical Framework for Relation Extraction with Reinforcement Learning},
  author={Takanobu, Ryuichi and Zhang, Tianyang and Liu, Jiexi and Huang, Minlie},
  booktitle={AAAI},
  pages={7072--7079},
  year={2019}
}

Data

NYT10 is originally released by the paper "Modeling relations and their mentions without labeled text." Download , while NYT11 is provided by the paper "CoType: Joint Extraction of Typed Entities and Relations with Knowledge Bases" Download

The preprocessed dataset that we used for relation extraction are provided in corresponding subdirectories under HRL-RE/data.

Each relational triple is formatted as follows:

• rtext : relation type
• em1 : source entity mention
• em2 : target entity mention
• tags : the proposed entity annotation scheme for the sentence
  • 0 : $O$ non-entity
  • 1 : $S_I$ inside of a source entity
  • 2 : $T_I$ inside of a target entity
  • 3 : $O_I$ inside of not-concerned entity
  • 4 : $S_B$ begin of a source entity
  • 5 : $T_B$ begin of a target entity
  • 6 : $O_B$ begin of not-concerned entity

Run

Command

cd code
python main.py {--[option1]=[value1] --[option2]=[value2] ... }

Change the corresponding options to set hyper-parameters:

parser.add_argument('--logfile', type=str, default='HRL', help="Filename of log file")
parser.add_argument('--lr', type=float, default=0.0001, help="Learning rate")
parser.add_argument('--epochPRE', type=int, default=15, help="Number of epoch on pretraining")
parser.add_argument('--epochRL', type=int, default=10, help="Number of epoch on training with RL")
parser.add_argument('--dim', type=int, default=300, help="Dimension of hidden layer")
parser.add_argument('--statedim', type=int, default=300, help="Dimension of state")
parser.add_argument('--batchsize', type=int, default=16, help="Batch size on training")
parser.add_argument('--batchsize_test', type=int, default=64, help="Batch size on testing")
parser.add_argument('--print_per_batch', type=int, default=200, help="Print results every XXX batches")
parser.add_argument('--sampleround', type=int, default=5, help="Sample round in RL")
parser.add_argument('--numprocess', type=int, default=4, help="Number of process")
parser.add_argument('--start', type=str, default='', help="Directory to load model")
parser.add_argument('--test', type=bool, default=False, help="Set to True to inference")
parser.add_argument('--pretrain', type=bool, default=False, help="Set to True to pretrain")
parser.add_argument('--datapath', type=str, default='../data/NYT10/', help="Data directory")
parser.add_argument('--testfile', type=str, default='test', help="Filename of test file")

NOTE: please start with pretraining, e.g.

python main.py --datapath ../data/NYT10/ --pretrain True

then,

python main.py --lr 2e-5 --datapath ../data/NYT10/ --start checkpoints/model_HRL_10

Requirements

• python 3
• pytorch = 0.3