Joined Type Length Encoding for Nested Named Entity Recognition

In this article, we propose a new encoding scheme for named entity recognition (NER) called Joined Type-Length encoding (JoinedTL). Unlike most existing named entity encoding schemes, which focus on flat entities, JoinedTL can label nested named entities in a single sequence. JoinedTL uses a packed encoding to represent both type and span of a named entity, which not only results in less tagged tokens compared to existing encoding schemes, but also enables it to support nested NER. We evaluate the effectiveness of JoinedTL for nested NER on three nested NER datasets: GENIA in English, GermEval in German, and PerNest, our newly created nested NER dataset in Persian. We apply CharLSTM+WordLSTM+CRF, a three-layer sequence tagging model on three datasets encoded using JoinedTL and two existing nested NE encoding schemes, i.e., JoinedBIO and JoinedBILOU. Our experiment results show that CharLSTM+WordLSTM+CRF trained with JoinedTL encoded datasets can achieve competitive F1 scores as the ones trained with datasets encoded by two other encodings, but with 27%–48% less tagged tokens. To leverage the power of three different encodings, i.e., JoinedTL, JoinedBIO, and JoinedBILOU, we propose an encoding-based ensemble method for nested NER. Evaluation results show that the ensemble method achieves higher F1 scores on all datasets than the three models each trained using one of the three encodings. By using nested NE encodings including JoinedTL with CharLSTM+WordLSTM+CRF, we establish new state-of-the-art performance with an F1 score of 83.7 on PerNest, 74.9 on GENIA, and 70.5 on GermEval, surpassing two recent neural models specially designed for nested NER.

TBCOV: Two Billion Multilingual COVID-19 Tweets with Sentiment, Entity, Geo, and Gender Labels

As the world struggles with several compounded challenges caused by the COVID-19 pandemic in the health, economic, and social domains, timely access to disaggregated national and sub-national data are important to understand the emergent situation but it is difficult to obtain. The widespread usage of social networking sites, especially during mass convergence events, such as health emergencies, provides instant access to citizen-generated data offering rich information about public opinions, sentiments, and situational updates useful for authorities to gain insights. We offer a large-scale social sensing dataset comprising two billion multilingual tweets posted from 218 countries by 87 million users in 67 languages. We used state-of-the-art machine learning models to enrich the data with sentiment labels and named-entities. Additionally, a gender identification approach is proposed to segregate user gender. Furthermore, a geolocalization approach is devised to geotag tweets at country, state, county, and city granularities, enabling a myriad of data analysis tasks to understand real-world issues at national and sub-national levels. We believe this multilingual data with broader geographical and longer temporal coverage will be a cornerstone for researchers to study impacts of the ongoing global health catastrophe and to manage adverse consequences related to people’s health, livelihood, and social well-being.

Training set augmentation in training neural- network language model for ontology population

This paper is a continuation of the research focused on solving the problem of ontology population using training on an automatically generated training set and the subsequent use of a neural-network language model for analyzing texts in order to discover new concepts to add to the ontology. The article is devoted to the text data augmentation - increasing the size of the training set by modification of its samples. Along with this, a solution to the problem of clarifying concepts (i.e. adjusting their boundaries in sentences), which were found during the automatic formation of the training set, is considered. A brief overview of existing approaches to text data augmentation, as well as approaches to extracting so-called nested named entities (nested NER), is presented. A procedure is proposed for clarifying the boundaries of the discovered concepts of the training set and its augmentation for subsequent training a neural-network language model in order to identify new concepts of ontology in the domain texts. The results of the experimental evaluation of the trained model and the main directions of further research are considered.

Indirectly Named Entity Recognition

We define here indirectly named entities, as a term to denote multiword expressions referring to known named entities by means of periphrasis. While named entity recognition is a classical task in natural language processing, little attention has been paid to indirectly named entities and their treatment. In this paper, we try to address this gap, describing issues related to the detection and understanding of indirectly named entities in texts. We introduce a proof of concept for retrieving both lexicalised and non-lexicalised indirectly named entities in French texts. We also show example cases where this proof of concept is applied, and discuss future perspectives. We have initiated the creation of a first lexicon of 712 indirectly named entity entries that is available for future research.

MELHISSA: a multilingual entity linking architecture for historical press articles

Digital libraries have a key role in cultural heritage as they provide access to our culture and history by indexing books and historical documents (newspapers and letters). Digital libraries use natural language processing (NLP) tools to process these documents and enrich them with meta-information, such as named entities. Despite recent advances in these NLP models, most of them are built for specific languages and contemporary documents that are not optimized for handling historical material that may for instance contain language variations and optical character recognition (OCR) errors. In this work, we focused on the entity linking (EL) task that is fundamental to the indexation of documents in digital libraries. We developed a Multilingual Entity Linking architecture for HIstorical preSS Articles that is composed of multilingual analysis, OCR correction, and filter analysis to alleviate the impact of historical documents in the EL task. The source code is publicly available. Experimentation has been done over two historical documents covering five European languages (English, Finnish, French, German, and Swedish). Results have shown that our system improved the global performance for all languages and datasets by achieving an F-score@1 of up to 0.681 and an F-score@5 of up to 0.787.

A comparative analysis of text embedding approach to extract named entities in Portuguese legal documents

A petição inicial é um dos componentes mais importantes de um processo civil, de modo que a automatização da análise desses documentos pode diminuir o tempo necessário para que se cumpra a fase postulatória. O corpo de qualificação das partes, por sua vez, é a seção nesse documento onde são expostas as informações a respeito das entidades envolvidas no processo. Este artigo propõe o uso de técnicas de extração de entidades nomeadas no problema de identificação e extração de informações em petições iniciais. Para tal, foi produzida uma base de dados composta por corpos de qualificação das partes de petições iniciais extraídas de processos advindos de tribunais brasileiros. Foram treinados, avaliados e comparados sete modelos BiLSTM-CRF com combinações distintas de representações vetoriais de palavras, a fim de se investigar seus efeitos na performance de um algoritmo com essa arquitetura e, dessa forma, aprimorar o reconhecimento de entidades jurídicas em textos legais. Ao contrário de outros trabalhos baseados em BiLSTM-CRF para tarefas de NER no domínio jurídico, esta pesquisa dá ênfase não às arquiteturas empregadas, mas sim aos métodos de representação de texto usados. Os experimentos executados com o corpus desenvolvido mostram que o empilhamento de incorporações de caracteres, palavras e pooled FLAIR embeddings é a combinação preferível para extrair-se o melhor desempenho possível de modelos híbridos BiLSTM-CRF.

Unsupervised Named Entity Recognition for Hi-Tech Domain

This paper presents named entity recognition as a multi-answer QA task combined with contextual natural-language-inference based noise reduction. This method allows us to use pre-trained models that have been trained for certain downstream tasks to generate unsupervised data, reducing the need for manual annotation to create named entity tags with tokens. For each entity, we provide a unique context, such as entity types, definitions, questions and a few empirical rules along with the target text to train a named entity model for the domain of our interest. This formulation (a) allows the system to jointly learn NER-specific features from the datasets provided, and (b) can extract multiple NER-specific features, thereby boosting the performance of existing NER models (c) provides business-contextualized definitions to reduce ambiguity among similar entities. We conducted numerous tests to determine the quality of the created data, and we find that this method of data generation allows us to obtain clean, noise-free data with minimal effort and time. This approach has been demonstrated to be successful in extracting named entities, which are then used in subsequent components.

Ensemble of Deep Masked Language Models for Effective Named Entity Recognition in Health and Life Science Corpora

The health and life science domains are well known for their wealth of named entities found in large free text corpora, such as scientific literature and electronic health records. To unlock the value of such corpora, named entity recognition (NER) methods are proposed. Inspired by the success of transformer-based pretrained models for NER, we assess how individual and ensemble of deep masked language models perform across corpora of different health and life science domains—biology, chemistry, and medicine—available in different languages—English and French. Individual deep masked language models, pretrained on external corpora, are fined-tuned on task-specific domain and language corpora and ensembled using classical majority voting strategies. Experiments show statistically significant improvement of the ensemble models over an individual BERT-based baseline model, with an overall best performance of 77% macro F1-score. We further perform a detailed analysis of the ensemble results and show how their effectiveness changes according to entity properties, such as length, corpus frequency, and annotation consistency. The results suggest that the ensembles of deep masked language models are an effective strategy for tackling NER across corpora from the health and life science domains.

NERO: a biomedical named-entity (recognition) ontology with a large, annotated corpus reveals meaningful associations through text embedding

Machine reading (MR) is essential for unlocking valuable knowledge contained in millions of existing biomedical documents. Over the last two decades1,2, the most dramatic advances in MR have followed in the wake of critical corpus development3. Large, well-annotated corpora have been associated with punctuated advances in MR methodology and automated knowledge extraction systems in the same way that ImageNet4 was fundamental for developing machine vision techniques. This study contributes six components to an advanced, named entity analysis tool for biomedicine: (a) a new, Named Entity Recognition Ontology (NERO) developed specifically for describing textual entities in biomedical texts, which accounts for diverse levels of ambiguity, bridging the scientific sublanguages of molecular biology, genetics, biochemistry, and medicine; (b) detailed guidelines for human experts annotating hundreds of named entity classes; (c) pictographs for all named entities, to simplify the burden of annotation for curators; (d) an original, annotated corpus comprising 35,865 sentences, which encapsulate 190,679 named entities and 43,438 events connecting two or more entities; (e) validated, off-the-shelf, named entity recognition (NER) automated extraction, and; (f) embedding models that demonstrate the promise of biomedical associations embedded within this corpus.