thunlp / mnre Goto Github PK

The code and data for ACL2017 paper "Neural Relation Extraction with Multi-lingual Attention"

License: MIT License

C++ 99.74% Makefile 0.26%

mnre's Introduction

Neural Relation Extraction with Multi-lingual Attention (MNRE)

Neural relation extraction aims to extract relations from plain text with neural models, which has been the state-of-the-art methods for relation extraction. In this project, we provide our implementations of CNN [Zeng et al., 2014] and PCNN [Zeng et al.,2015] and their extended version with multi-lingual sentence-level attention scheme [Lin et al., 2017] .

Data

We provide the dataset we used for the task relation extraction in (https://pan.baidu.com/s/1dF26l93). We preprocess the original data to make it satisfy the input format of our codes.

Pre-Trained English Word Vectors are learned from New York Times Annotated Corpus (LDC Data LDC2008T19), which should be obtained from LDC (https://catalog.ldc.upenn.edu/LDC2008T19).

Pre-Trained Chinese Word Vectors are learned from Chinese Baidu Baike (https://baike.baidu.com/).

To run our code, the dataset should be put in the folder data/ using the following format, containing six files

train_en.txt / train_zh.txt: training file, format (wikidata_qid_e1, wikidata_qid_e2, e1_name, e2_name, relation, sentence).
valid_en.txt / valid_zh.txt: validation file, same format as train.txt
test_en.txt / test_zh.txt: test file, same format as train.txt.
entity2id.txt: all entities and corresponding ids, one per line.
relation2id.txt: all relations and corresponding ids, one per line.
vec_en.bin, vec_zh.bin: the pre-train word embedding file

Codes

The source codes of various methods are put in the folders src/.

Compile

Just type "make" in the folder src/.

Train

For training, you need to type the following command in each model folder:

./train

The training model file will be saved in folder out/ .

Test

For testing, you need to type the following command in each model folder:

./test

The testing result which reports the precision/recall curve will be shown in pr.txt.

Cite

If you use the code, please cite the following paper:

[Lin et al., 2017] Yankai Lin, Zhiyuan Liu, and Maosong Sun. Neural Relation Extraction with Multi-lingual Attention. In Proceedings of ACL.[pdf]

Reference

[Zeng et al., 2014] Daojian Zeng, Kang Liu, Siwei Lai, Guangyou Zhou, and Jun Zhao. Relation classification via convolutional deep neural network. In Proceedings of COLING.

[Zeng et al.,2015] Daojian Zeng,Kang Liu,Yubo Chen,and Jun Zhao. Distant supervision for relation extraction via piecewise convolutional neural networks. In Proceedings of EMNLP.

[Lin et al., 2016] Yankai Lin, Shiqi Shen, Zhiyuan Liu, Huanbo Luan, and Maosong Sun. Neural Relation Extraction with Selective Attention over Instances. In Proceedings of ACL.[pdf]

mnre's People

Contributors

Stargazers

Watchers

Forkers

waterzxj jasonhoou aabbcc0812206523 huokedu li10141110 yyaghoobzadeh stevenlol jx57 sxdkxgwan mqrshiyan zhaohuiqiang readyteresa baylee001 sandy4321 austin-w

mnre's Issues

a new issue

what is mean the code "if(res !=res)",in my opinion,program will never operate it in train.cpp of cnn file!

An error “fprintf_chk.c: No such file or directory” occurs

您好，
When I execute the original code under Ubuntu 14.04., the error occurs:

Segmentation fault (core dumped)

Then, I tried to debug it with gdb, and got the following error:

The program received signal SIGSEGV, Segmentation fault.
___fprintf_chk (fp=0x0, flag=1, format=0x4155c6 "%d\n") at fprintf_chk.c:30
30      fprintf_chk.c: No such file or directory.

I have googled a lot, and find no solution, could you give me some advice to resolve that.

Thanks so much.

Kind Regards.

Segmentation fault (core dumped)

gpuws@gpuws32g:/ub16_prj/MNRE/src/MNRE_CNN$ ./train
1 4 3 1 5 1 4 0 3 1 Init Begin.
Begin_init_para
relationTotal: 176
End_init_para
Begin read data of language: zh
wordTotal= 4756241
Word dimension= 50
240
End read data of language: zh
Begin read data of language: en
wordTotal= 1067259
Word dimension= 50
240
End read data of language: en
540098 537992
Init End.
959935
npoch: 187/1874
time(s): 394.44
score: -135516 95744
npoch: 374/1874
time(s): 395.614
score: -111949 191488
npoch: 561/1874
time(s): 421.19
score: -108274 287232
npoch: 748/1874
time(s): 497.307
score: -104558 382976
npoch: 935/1874
time(s): 799.708
score: -100079 478720
npoch: 1122/1874
time(s): 727.027
score: -94617.7 574464
npoch: 1309/1874
time(s): 778.266
score: -88565.3 670208
npoch: 1496/1874
time(s): 661.395
score: -82241.4 765952
npoch: 1683/1874
time(s): 634.374
score: -77362.2 861696
npoch: 1870/1874
time(s): 585.224
score: -73005.4 957440
Total Score: -977725.375000
test
version=1
tot: 47638 537992
94148600
precision: 0 recall: 0
precision: 0.0738523 recall: 0.000776691
precision: 0.0849151 recall: 0.00178429
precision: 0.0852765 recall: 0.00268693
precision: 0.0894553 recall: 0.0037575
precision: 0.0903639 recall: 0.00474411
precision: 0.0943019 recall: 0.00594064
precision: 0.0945444 recall: 0.00694823
precision: 0.0977256 recall: 0.00820773
precision: 0.0986447 recall: 0.00932029
precision: 0.09978 recall: 0.0104748
precision: 0.100709 recall: 0.0116294
precision: 0.10215 recall: 0.0128679
precision: 0.1026 recall: 0.0140014
precision: 0.103414 recall: 0.015198
precision: 0.103586 recall: 0.0163105
precision: 0.103487 recall: 0.0173811
precision: 0.102576 recall: 0.0183047
precision: 0.101989 recall: 0.0192703
precision: 0.102726 recall: 0.0204878
1
Segmentation fault (core dumped)
gpuws@gpuws32g:/ub16_prj/MNRE/src/MNRE_CNN$

关于PR曲线

你好，
能否上传一份关于这篇论文Neural Relation Extraction with Multi-lingual Attention中汇报的P/R曲线的数据。
谢谢！

Segmentation fault

When I was doing the training step,It came a problem with "Segmentation fault".Have you met it before?