# Improving Chemical Named Entity Recognition in Patents with Contextualized Word Embeddings

arXiv: https://arxiv.org/abs/1907.02679 · HTML: https://mlsanity.com/paper/1907.02679/

## Key facts

- Cited by 9 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 5 Jul 2019.

## Built on by (most cited first, 9 of 9)

- [Predictive Chemistry Augmented with Text Retrieval](https://mlsanity.com/paper/2312.04881/) (2023)
- [MarkushGrapher: Joint Visual and Textual Recognition of Markush Structures](https://mlsanity.com/paper/2503.16096/) (2025)
- [Stress Test Evaluation of Biomedical Word Embeddings](https://mlsanity.com/paper/2107.11652/) (2021)
- [MarkushGrapher-2: End-to-end Multimodal Recognition of Chemical Structures](https://mlsanity.com/paper/2603.28550/) (2026)
- [Text to Insight: Accelerating Organic Materials Knowledge Extraction via Deep Learning](https://mlsanity.com/paper/2109.12758/) (2021)
- [An Intellectual Property Entity Recognition Method Based on Transformer and Technological Word Information](https://mlsanity.com/paper/2203.10717/) (2022)
- [Research on Intellectual Property Resource Profile and Evolution Law](https://mlsanity.com/paper/2204.06221/) (2022)
- [Stress Testing BERT Anaphora Resolution Models for Reaction Extraction in Chemical Patents](https://mlsanity.com/paper/2306.13379/) (2023)
- [Chemical Reaction Extraction from Long Patent Documents](https://mlsanity.com/paper/2407.15124/) (2024)

## Abstract (arXiv)

Chemical patents are an important resource for chemical information. However, few chemical Named Entity Recognition (NER) systems have been evaluated on patent documents, due in part to their structural and linguistic complexity. In this paper, we explore the NER performance of a BiLSTM-CRF model utilising pre-trained word embeddings, character-level word representations and contextualized ELMo word representations for chemical patents. We compare word embeddings pre-trained on biomedical and chemical patent corpora. The effect of tokenizers optimized for the chemical domain on NER performance in chemical patents is also explored. The results on two patent corpora show that contextualized word representations generated from ELMo substantially improve chemical NER performance w.r.t. the current state-of-the-art. We also show that domain-specific resources such as word embeddings trained on chemical patents and chemical-specific tokenizers have a positive impact on NER performance.