[SCRIPT] Added script for named entity recognition experiments

[vanewu]

Description

This module will contain several experiments based on deep learning for named entity recognition. Currently, the model of CharCNN+BiLSTM+CRF has been added, which was proposed by Xuezhe Ma et al. 2016.

Implemented

1. Data preprocessing.
2. Linear chain conditional random fields.
3. CharCNN+BiLSTM+CRF model.
4. Model training and evaluation with conll2003evaluate.sh.
5. Gradient clip and learning rate decay.

Checklist

Essentials

• [x] PR's title starts with a category (e.g. [BUGFIX], [MODEL], [TUTORIAL], [FEATURE], [DOC], etc)
• [ ] Changes are complete (i.e. I finished coding on this PR)
• [x] All changes have test coverage
• [x] Code is well-documented

Changes

Added CharCNN+BiLSTM+CRF model to scripts/named_entity_recognition.

Comments

[codecov[bot]]

Codecov Report

Merging #466 into master will decrease coverage by 3.67%. The diff coverage is 0%.

@@            Coverage Diff             @@
##           master     #466      +/-   ##
==========================================
- Coverage   71.31%   67.63%   -3.68%     
==========================================
  Files         118      123       +5     
  Lines        9889    10434     +545     
==========================================
+ Hits         7052     7057       +5     
- Misses       2837     3377     +540

Flag	Coverage Δ
#PR466	67.63% <0%> (?)
#master	?
#notserial	44.01% <0%> (-2.56%)	:arrow_down:
#py2	67.34% <0%> (-3.73%)	:arrow_down:
#py3	67.5% <0%> (-3.67%)	:arrow_down:
#serial	54.15% <0%> (-3.02%)	:arrow_down:

[codecov[bot]]

Codecov Report

Merging #466 into master will decrease coverage by 7.07%. The diff coverage is 34.34%.

@@            Coverage Diff             @@
##           master     #466      +/-   ##
==========================================
- Coverage   69.91%   62.84%   -7.08%     
==========================================
  Files         122      151      +29     
  Lines       10461    14003    +3542     
==========================================
+ Hits         7314     8800    +1486     
- Misses       3147     5203    +2056

Flag	Coverage Δ
#PR435	63.87% <34.97%> (?)
#PR466	64.63% <34.97%> (-25.06%)	:arrow_down:
#PR588	89.46% <81.54%> (?)
#PR612	63.05% <34.97%> (?)
#PR639	63.65% <34.94%> (?)
#PR648	63.82% <34.97%> (?)
#master	63.78% <34.97%> (-6.13%)	:arrow_down:
#notserial	39.27% <20.49%> (-7.3%)	:arrow_down:
#py2	63.64% <34.18%> (-6.05%)	:arrow_down:
#py3	62.72% <34.34%> (-7%)	:arrow_down:
#serial	49.4% <26.88%> (-5.15%)	:arrow_down:

[mli]

Job PR-466/6 is complete. Docs are uploaded to http://gluon-nlp-staging.s3-accelerate.dualstack.amazonaws.com/PR-466/6/index.html

[mli]

Job PR-466/11 is complete. Docs are uploaded to http://gluon-nlp-staging.s3-accelerate.dualstack.amazonaws.com/PR-466/11/index.html

[mli]

Job PR-466/16 is complete. Docs are uploaded to http://gluon-nlp-staging.s3-accelerate.dualstack.amazonaws.com/PR-466/16/index.html

[mli]

Job PR-466/18 is complete. Docs are uploaded to http://gluon-nlp-staging.s3-accelerate.dualstack.amazonaws.com/PR-466/18/index.html

[mli]

Job PR-466/20 is complete. Docs are uploaded to http://gluon-nlp-staging.s3-accelerate.dualstack.amazonaws.com/PR-466/20/index.html

[mli]

Job PR-466/22 is complete. Docs are uploaded to http://gluon-nlp-staging.s3-accelerate.dualstack.amazonaws.com/PR-466/22/index.html

[mli]

Job PR-466/22 is complete. Docs are uploaded to http://gluon-nlp-staging.s3-accelerate.dualstack.amazonaws.com/PR-466/22/index.html

[eric-haibin-lin]

@kenjewu any update on this?