# Robust Lexical Features for Improved Neural Network Named-Entity Recognition

arXiv: https://arxiv.org/abs/1806.03489 · HTML: https://mlsanity.com/paper/1806.03489/

## Key facts

- Cited by 12 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 9 Jun 2018.

## Built on by (most cited first, 12 of 12)

- [SpanNER: Named Entity Re-/Recognition as Span Prediction](https://mlsanity.com/paper/2106.00641/) (2021)
- [Better Feature Integration for Named Entity Recognition](https://mlsanity.com/paper/2104.05316/) (2021)
- [Global Pointer: Novel Efficient Span-based Approach for Named Entity Recognition](https://mlsanity.com/paper/2208.03054/) (2022)
- [Recurrent Attention Networks for Long-text Modeling](https://mlsanity.com/paper/2306.06843/) (2023)
- [What do we Really Know about State of the Art NER?](https://mlsanity.com/paper/2205.00034/) (2022)
- [Doc2Dict: Information Extraction as Text Generation](https://mlsanity.com/paper/2105.07510/) (2021)
- [Larger-Context Tagging: When and Why Does It Work?](https://mlsanity.com/paper/2104.04434/) (2021)
- [Korean Named Entity Recognition Based on Language-Specific Features](https://mlsanity.com/paper/2305.06330/) (2023)
- [Unleashing the True Potential of Sequence-to-Sequence Models for Sequence Tagging and Structure Parsing](https://mlsanity.com/paper/2302.02275/) (2023)
- [A Global Context Mechanism for Sequence Labeling](https://mlsanity.com/paper/2305.19928/) (2023)
- [Enhanced E-Commerce Attribute Extraction: Innovating with Decorative Relation Correction and LLAMA 2.0-Based Annotation](https://mlsanity.com/paper/2312.06684/) (2023)
- [Cross-Lingual News Event Correlation for Stock Market Trend Prediction](https://mlsanity.com/paper/2410.00024/) (2024)

## Abstract (arXiv)

Neural network approaches to Named-Entity Recognition reduce the need for carefully hand-crafted features. While some features do remain in state-of-the-art systems, lexical features have been mostly discarded, with the exception of gazetteers. In this work, we show that this is unfair: lexical features are actually quite useful. We propose to embed words and entity types into a low-dimensional vector space we train from annotated data produced by distant supervision thanks to Wikipedia. From this, we compute - offline - a feature vector representing each word. When used with a vanilla recurrent neural network model, this representation yields substantial improvements. We establish a new state-of-the-art F1 score of 87.95 on ONTONOTES 5.0, while matching state-of-the-art performance with a F1 score of 91.73 on the over-studied CONLL-2003 dataset.