LSTMVoter: chemical named entity recognition using a conglomerate of sequence labeling tools

Abstract

Background Chemical and biomedical named entity recognition (NER) is an essential preprocessing task in natural language processing . The identification and extraction of named entities from scientific articles is also attracting increasing interest in many scientific disciplines. Locating chemical named entities in the literature is an essential step in chemical text mining pipelines for identifying chemical mentions, their properties, and relations as discussed in the literature. In this work, we describe an approach to the BioCreative V.5 challenge regarding the recognition and classification of chemical named entities. For this purpose, we transform the task of NER into a sequence labeling problem. We present a series of sequence labeling systems that we used, adapted and optimized in our experiments for solving this task. To this end, we experiment with hyperparameter optimization. Finally, we present LSTMVoter, a two-stage application of recurrent neural network s that integrates the optimized sequence labelers from our study into a single ensemble classifier. Results We introduce LSTMVoter, a bidirectional long short-term memory (LSTM) tagger that utilizes a conditional random field layer in conjunction with attention-based feature modeling. Our approach explores information about features that is modeled by means of an attention mechanism. LSTMVoter outperforms each extractor integrated by it in a series of experiments. On the BioCreative IV chemical compound and drug name recognition (CHEMDNER) corpus, LSTMVoter achieves an F1-score of 90.04%; on the BioCreative V.5 chemical entity mention in patents corpus, it achieves an F1-score of 89.01%. Availability and implementation Data and code are available at https://github.com/texttechnologylab/LSTMVoter .

Automated Extraction of Information From Texts of Scientific Publications: Insights Into HIV Treatment Strategies

“…There are methods aimed at NER that have been developing during the last years (Kaewphan et al, 2018 ; Korvigo et al, 2018 ; Hemati and Mehler, 2019 ; Hong and Lee, 2020 ; Huang et al, 2020 ; Kilicoglu et al, 2020 ). Most of them are based on algorithms for NER related either to chemicals or biological objects.…”

ChEMU 2020: Natural Language Processing Methods Are Effective for Information Extraction From Chemical Patents

“…Early techniques for chemical text mining, such as dictionary-based methods (Rebholz-Schuhmann et al, 2007 ; Hettne et al, 2009 ; Akhondi et al, 2016 ) and grammar-based methods (Narayanaswamy et al, 2002 ; Liu et al, 2012 ; Akhondi et al, 2015 ), heavily rely on expert knowledge in the chemical domain. Recently, machine learning-based techniques have reported state-of-the-art effectiveness in chemical text mining (Hemati and Mehler, 2019 ; Zhai et al, 2019 ). However, such techniques require a large amount of annotated text data, which still remains limited.…”

Deep neural model with enhanced embeddings for pharmaceutical and chemical entities recognition in Spanish clinical text

“…We used the PyMedTermino library (Lamy et al, 2015) for concept indexing. A full-text search with the Levenshtein distance algorithm (Miller et al, 2009) was applied in a first instance for concept indexing and fuzzy search with threshold using FuzzyDict implementation (Hemati and Mehler, 2019) as a second approach for concepts not found by partial matching. The FastText model uses a combination of various subcomponents to produce high-quality embeddings.…”

“…One of the most effective methods is Conditional Random Fields (CRF) (Lafferty et al, 2001) since CRF is one of the most reliable sequence labeling methods. Recently, deep learning-based methods have also demonstrated state-of-the-art performance for English (Hemati and Mehler, 2019;Pérez-Pérez et al, 2017;Suárez-Paniagua et al, 2019) texts by automatically learning relevant patterns from corpora, which allows language and domain independence. However, until now, to the best of our knowledge, there is only one work that addresses the generation of Spanish biomedical word embeddings (Armengol-Estapé Jordi, 2019;Soares et al, 2019).…”