Chinese NER Using Lattice LSTM
arXiv:1805.02023
Abstract
We investigate a lattice-structured LSTM model for Chinese NER, which encodes a sequence of input characters as well as all potential words that match a lexicon. Compared with character-based methods, our model explicitly leverages word and word sequence information. Compared with word-based methods, lattice LSTM does not suffer from segmentation errors. Gated recurrent cells allow our model to choose the most relevant characters and words from a sentence for better NER results. Experiments on various datasets show that lattice LSTM outperforms both word-based and character-based LSTM baselines, achieving the best results.
arXiv commentAccepted at ACL 2018 as Long paper
References in corpus (5)
- Empower Sequence Labeling with Task-Aware Neural Language Model
- Lattice-Based Recurrent Neural Network Encoders for Neural Machine Translation
- Adversarial Multi-Criteria Learning for Chinese Word Segmentation
- Improving Named Entity Recognition for Chinese Social Media with Word Segmentation Representation Learning
- Neural Lattice-to-Sequence Models for Uncertain Inputs
Cited by in corpus (12)
- Joint Extraction of Entities and Relations Based on a Novel Decomposition Strategy
- A Unified MRC Framework for Named Entity Recognition
- Dice Loss for Data-imbalanced NLP Tasks
- Interpreting Deep Learning Models in Natural Language Processing: A Review
- Self-attention-based BiGRU and capsule network for named entity recognition
- Generating Adversarial Examples in Chinese Texts Using Sentence-Pieces
- Neural Entity Reasoner for Global Consistency in NER
- Lattice CNNs for Matching Based Chinese Question Answering
- Learning Task-specific Representation for Novel Words in Sequence Labeling
- Incorporating Uncertain Segmentation Information into Chinese NER for Social Media Text
- Query-Based Named Entity Recognition
- Improving NER's Performance with Massive financial corpus