Sequence Labeling: A Practical Approach
arXiv:1808.03926
Abstract
We take a practical approach to solving sequence labeling problem assuming unavailability of domain expertise and scarcity of informational and computational resources. To this end, we utilize a universal end-to-end Bi-LSTM-based neural sequence labeling model applicable to a wide range of NLP tasks and languages. The model combines morphological, semantic, and structural cues extracted from data to arrive at informed predictions. The model's performance is evaluated on eight benchmark datasets (covering three tasks: POS-tagging, NER, and Chunking, and four languages: English, German, Dutch, and Spanish). We observe state-of-the-art results on four of them: CoNLL-2012 (English NER), CoNLL-2002 (Dutch NER), GermEval 2014 (German NER), Tiger Corpus (German POS-tagging), and competitive performance on the rest.
arXiv commentFor the source code and detailed experimental results, see http://github.com/aakhundov/sequence-labeling
References in corpus (7)
- Adam: A Method for Stochastic Optimization
- Recurrent Neural Network Regularization
- Introduction to the CoNLL-2002 Shared Task: Language-Independent Named Entity Recognition
- On the Properties of Neural Machine Translation: Encoder-Decoder Approaches
- Multi-Task Cross-Lingual Sequence Tagging from Scratch
- Robust Multilingual Named Entity Recognition with Shallow Semi-Supervised Features
- Neural Models for Sequence Chunking
Cited by in corpus (7)
- TENER: Adapting Transformer Encoder for Named Entity Recognition
- Few-shot classification in Named Entity Recognition Task
- Phraseformer: Multimodal Key-phrase Extraction using Transformer and Graph Embedding
- Multi-Scale Self-Attention for Text Classification
- Aspect-Based Argument Mining
- Active Learning for Argument Mining: A Practical Approach
- Method and Dataset Entity Mining in Scientific Literature: A CNN + Bi-LSTM Model with Self-attention