Awesome

Active Learning for Entity Recognition

Requirements

python 2.7 DynetVersion commit 284838815ece9297a7100cc43035e1ea1b133a5

Data

In the data/, create a directory per language as shown for data/Spanish. Download the CoNLL train/dev/test NER datasets for that language here. To acquire LDC datasets, please get the required access.

For storing the trained models, create directory saved_models in the parent folder.

Embeddings

Combine monolingual data acquired from Wikipedia with the plain text extracted from the labeled data. Train 100-d Glove embeddings

Active Learning Simulation

The best NER performance was obtained using fine-tuning training scheme. The scripts below runs simulation active learning runs for different active learning strategies: cd commands

ETAL + Partial-CRF + CT (Proposed recipe) ./ETAL_PARTIAL_CRF_CT.sh
ETAL + Full-CRF + CT ./ETAL_FULL_CRF_CT.sh
CFEAL + Full-CRF + CT ./CFEAL_PARTIAL_CRF_CT.sh
SAL + CT ./SAL_CT.sh Things to note:

We load the vocabulary from the following path--aug_lang_train_path. Therefore, create a conll formatted file with dummy labels from the unlabeled text. For our experiments, we concatenated the transferred data with the unlabeled data (which was the entire training dataset) into a single conll formatted file. The conll format is a tab separated two-column format as shown below:

El O grupo O

The LDC NER label set differ from the CoNLL label set by one tag. Therefore, add --misc to the argument set when running any experiments on CoNLL data. The label set has been hard-coded in the data_loaders/data_loader.py file.

Cross-Lingual Transferred Data

We used the model proposed by (Xie et al. 2018) to get the cross-lingually transferred data from English. Please refer to their code here.

For the Fine-Tune training scheme, train a base NER model on the transferred model as follows:

MODEL_NAME="spanish_full_transfer_baseline"
python -u ../main.py \
    --dynet-seed 3278657 \
    --word_emb_dim 100 \
    --batch_size 10 \
    --model_name ${MODEL_NAME} \
    --lang es \
    --fixedVocab \
    --test_conll \
    --tot_epochs 1000 \
--aug_lang_train_path $DATA/vocab.conll \
    --init_lr 0.015 \
    --valid_freq 1300 \
    --misc \
    --pretrain_emb_path $DATA/esp.vec \
    --dev_path $DATA/esp.dev \
    --test_path $DATA/esp.test \
    --train_path $DIR/transferred_data.conll  2>&1 | tee ${MODEL_NAME}.log

References

If you make use of this software for research purposes, we will appreciate citing the following:

@inproceedings{chaudhary19emnlp,
    title = {A Little Annotation does a Lot of Good: A Study in Bootstrapping Low-resource Named Entity Recognizers},
    author = {Aditi Chaudhary and Jiateng Xie and Zaid Sheikh and Graham Neubig and Jaime Carbonell},
    booktitle = {Conference on Empirical Methods in Natural Language Processing (EMNLP)},
    address = {Hong Kong},
    month = {November},
    url = {http://arxiv.org/abs/1908.08983},
    year = {2019}
}

Contact

For any issues, please feel free to reach out to aschaudh@andrew.cmu.edu.