Dataset Information

Next generation community assessment of biomedical entity recognition web servers: metrics, performance, interoperability aspects of BeCalm.

ABSTRACT:

Background

Shared tasks and community challenges represent key instruments to promote research, collaboration and determine the state of the art of biomedical and chemical text mining technologies. Traditionally, such tasks relied on the comparison of automatically generated results against a so-called Gold Standard dataset of manually labelled textual data, regardless of efficiency and robustness of the underlying implementations. Due to the rapid growth of unstructured data collections, including patent databases and particularly the scientific literature, there is a pressing need to generate, assess and expose robust big data text mining solutions to semantically enrich documents in real time. To address this pressing need, a novel track called "Technical interoperability and performance of annotation servers" was launched under the umbrella of the BioCreative text mining evaluation effort. The aim of this track was to enable the continuous assessment of technical aspects of text annotation web servers, specifically of online biomedical named entity recognition systems of interest for medicinal chemistry applications.

Results

A total of 15 out of 26 registered teams successfully implemented online annotation servers. They returned predictions during a two-month period in predefined formats and were evaluated through the BeCalm evaluation platform, specifically developed for this track. The track encompassed three levels of evaluation, i.e. data format considerations, technical metrics and functional specifications. Participating annotation servers were implemented in seven different programming languages and covered 12 general entity types. The continuous evaluation of server responses accounted for testing periods of low activity and moderate to high activity, encompassing overall 4,092,502 requests from three different document provider settings. The median response time was below 3.74 s, with a median of 10 annotations/document. Most of the servers showed great reliability and stability, being able to process over 100,000 requests in a 5-day period.

Conclusions

The presented track was a novel experimental task that systematically evaluated the technical performance aspects of online entity recognition systems. It raised the interest of a significant number of participants. Future editions of the competition will address the ability to process documents in bulk as well as to annotate full-text documents.

SUBMITTER: Perez-Perez M

PROVIDER: S-EPMC6591930 | biostudies-literature |

REPOSITORIES: biostudies-literature

ACCESS DATA

Similar Datasets

Project description:MotivationThe explosive increase of biomedical literature has made information extraction an increasingly important tool for biomedical research. A fundamental task is the recognition of biomedical named entities in text (BNER) such as genes/proteins, diseases and species. Recently, a domain-independent method based on deep learning and statistical word embeddings, called long short-term memory network-conditional random field (LSTM-CRF), has been shown to outperform state-of-the-art entity-specific BNER tools. However, this method is dependent on gold-standard corpora (GSCs) consisting of hand-labeled entities, which tend to be small but highly reliable. An alternative to GSCs are silver-standard corpora (SSCs), which are generated by harmonizing the annotations made by several automatic annotation systems. SSCs typically contain more noise than GSCs but have the advantage of containing many more training examples. Ideally, these corpora could be combined to achieve the benefits of both, which is an opportunity for transfer learning. In this work, we analyze to what extent transfer learning improves upon state-of-the-art results for BNER.ResultsWe demonstrate that transferring a deep neural network (DNN) trained on a large, noisy SSC to a smaller, but more reliable GSC significantly improves upon state-of-the-art results for BNER. Compared to a state-of-the-art baseline evaluated on 23 GSCs covering four different entity classes, transfer learning results in an average reduction in error of approximately 11%. We found transfer learning to be especially beneficial for target datasets with a small number of labels (approximately 6000 or less).Availability and implementationSource code for the LSTM-CRF is available at https://github.com/Franck-Dernoncourt/NeuroNER/ and links to the corpora are available at https://github.com/BaderLab/Transfer-Learning-BNER-Bioinformatics-2018/.Supplementary informationSupplementary data are available at Bioinformatics online.

Project description:Increased interest in the use of word embeddings, such as word representation, for biomedical named entity recognition (BioNER) has highlighted the need for evaluations that aid in selecting the best word embedding to be used. One common criterion for selecting a word embedding is the type of source from which it is generated; that is, general (e.g., Wikipedia, Common Crawl), or specific (e.g., biomedical literature). Using specific word embeddings for the BioNER task has been strongly recommended, considering that they have provided better coverage and semantic relationships among medical entities. To the best of our knowledge, most studies have focused on improving BioNER task performance by, on the one hand, combining several features extracted from the text (for instance, linguistic, morphological, character embedding, and word embedding itself) and, on the other, testing several state-of-the-art named entity recognition algorithms. The latter, however, do not pay great attention to the influence of the word embeddings, and do not facilitate observing their real impact on the BioNER task. For this reason, the present study evaluates three well-known NER algorithms (CRF, BiLSTM, BiLSTM-CRF) with respect to two corpora (DrugBank and MedLine) using two classic word embeddings, GloVe Common Crawl (of the general type) and Pyysalo PM + PMC (specific), as unique features. Furthermore, three contextualized word embeddings (ELMo, Pooled Flair, and Transformer) are compared in their general and specific versions. The aim is to determine whether general embeddings can perform better than specialized ones on the BioNER task. To this end, four experiments were designed. In the first, we set out to identify the combination of classic word embedding, NER algorithm, and corpus that results in the best performance. The second evaluated the effect of the size of the corpus on performance. The third assessed the semantic cohesiveness of the classic word embeddings and their correlation with respect to several gold standards; while the fourth evaluates the performance of general and specific contextualized word embeddings on the BioNER task. Results show that the classic general word embedding GloVe Common Crawl performed better in the DrugBank corpus, despite having less word coverage and a lower internal semantic relationship than the classic specific word embedding, Pyysalo PM + PMC; while in the contextualized word embeddings the best results are presented in the specific ones. We conclude, therefore, when using classic word embeddings as features on the BioNER task, the general ones could be considered a good option. On the other hand, when using contextualized word embeddings, the specific ones are the best option.

Project description:Motivation:Recognition of biomedical entities from scientific text is a critical component of natural language processing and automated information extraction platforms. Modern named entity recognition approaches rely heavily on supervised machine learning techniques, which are critically dependent on annotated training corpora. These approaches have been shown to perform well when trained and tested on the same source. However, in such scenario, the performance and evaluation of these models may be optimistic, as such models may not necessarily generalize to independent corpora, resulting in potential non-optimal entity recognition for large-scale tagging of widely diverse articles in databases such as PubMed. Results:Here we aggregated published corpora for the recognition of biomolecular entities (such as genes, RNA, proteins, variants, drugs and metabolites), identified entity class overlap and performed leave-corpus-out cross validation strategy to test the efficiency of existing models. We demonstrate that accuracies of models trained on individual corpora decrease substantially for recognition of the same biomolecular entity classes in independent corpora. This behavior is possibly due to limited generalizability of entity-class-related features captured by individual corpora (model 'overtraining') which we investigated further at the orthographic level, as well as potential annotation standard differences. We show that the combined use of multi-source training corpora results in overall more generalizable models for named entity recognition, while achieving comparable individual performance. By performing learning-curve-based power analysis we further identified that performance is often not limited by the quantity of the annotated data. Availability and implementation:Compiled primary and secondary sources of the aggregated corpora are available on: https://github.com/dterg/biomedical_corpora/wiki and https://bitbucket.org/iAnalytica/bioner. Supplementary information:Supplementary data are available at Bioinformatics online.