Chinese NER Using Lattice LSTM

TL;DR

This paper introduces a lattice-structured LSTM model for Chinese Named Entity Recognition that effectively combines character and word information, outperforming traditional methods by reducing segmentation errors and leveraging richer context.

Contribution

The paper presents a novel lattice LSTM model that integrates lexicon-matched words with character sequences for improved Chinese NER performance.

Findings

Lattice LSTM outperforms character-based models.

Lattice LSTM surpasses word-based models.

Model achieves state-of-the-art results on multiple datasets.

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.