mrgonext/vie-ner-lstm

LSTM model for Vietnamese Named Entity Recognition

Vietnamese Named Entity Recognition (vie-ner-lstm)

---

Code by Thai-Hoang Pham at Alt Inc.

1. Introduction

vie-ner-lstm is a fast-implementation of the system described in a paper The Importance of Automatic Syntactic Features in Vietnamese Named Entity Recognition. This system is used to recognize named entities in Vietnamese texts and written by Python 2.7. The architecture of this system is two bidirectional LSTM layers followed by a feed-forward neural network. Finally, the output sequence is predicted by a softmax function.

Our system achieved an F1 score of 92.05% on VLSP standard testset. The performance of our system with each feature set is described in the following table.

Word2vec	POS	Chunk	Regex	F1
				62.87%
x				74.02%
x	x			85.90%
x		x		86.79%
x			x	74.13%
x	x	x	x	92.05%

2. Installation

This software depends on NumPy, Keras. You must have them installed before using vie-ner-lstm.

The simple way to install them is using pip:

	# pip install -U numpy keras

3. Usage

3.1. Data

The input data's format of vie-ner-lstm follows VLSP 2016 campaign format. There are four columns in this dataset including of word, pos, chunk, and named entity. For details, see sample data in 'data' directory. The table below describes an example Vietnamese sentence in VLSP dataset.

Word	POS	Chunk	NER
Từ	E	B-PP	O
Singapore	NNP	B-NP	B-LOC
,	CH	O	O
chỉ	R	O	O
khoảng	N	B-NP	O
vài	L	B-NP	O
chục	M	B-NP	O
phút	Nu	B-NP	O
ngồi	V	B-VP	O
phà	N	B-NP	O
là	V	B-VP	O
dến	V	B-VP	O
được	R	O	O
Batam	NNP	B-NP	B-LOC
.	CH	O	O

To access the full dataset of VLSP 2016 campaign, you need to sign the user agreement of the VLSP consortium.

3.2. Command-line Usage

You can use vie-ner-lstm software by a following command:

	$ bash ner.sh

Arguments in ner.sh script:

• word_dir: path for word dictionary
• vector_dir: path for vector dictionary
• train_dir: path for training data
• dev_dir: path for development data
• test_dir: path for testing data
• num_lstm_layer: number of LSTM layers used in this system
• num_hidden_node: number of hidden nodes in a hidden LSTM layer
• dropout: dropout for input data (The float number between 0 and 1)
• batch_size: size of input batch for training this system.
• patience: number used for early stopping in training stage

Note: In the first time of running vie-ner-lstm, this system will automatically download word embeddings for Vietnamese from the internet. (It may take a long time because a size of this embedding set is about 1 GB). If the system cannot automatically download this embedding set, you can manually download it from here (vector, word) and put it into embedding directory.

4. References

Thai-Hoang Pham, Phuong Le-Hong, "The Importance of Automatic Syntactic Features in Vietnamese Named Entity Recognition" Proceedings of the 31th Pacific Asia Conference on Language, Information and Computation (PACLIC 31)

@inproceedings{Pham:2017,
  title={The Importance of Automatic Syntactic Features in Vietnamese Named Entity Recognition},
  author={Thai-Hoang Pham and Phuong Le-Hong},
  booktitle={Proceedings of the 31th Pacific Asia Conference on Language, Information and Computation},
  year={2017},
}

5. Contact

Thai-Hoang Pham < [email protected] >

Alt Inc, Hanoi, Vietnam