Using MEDLINE as a knowledge source for disambiguating abbreviations and acronyms in full-text biomedical journal articles

Simulation and annotation of global acronyms

MOTIVATION

Global acronyms are used in written text without their formal definitions. This makes it difficult to automatically interpret their sense as acronyms tend to be ambiguous. Supervised machine learning approaches to sense disambiguation require large training datasets. In clinical applications, large datasets are difficult to obtain due to patient privacy. Manual data annotation creates an additional bottleneck.

RESULTS

We proposed an approach to automatically modifying scientific abstracts to (i) simulate global acronym usage and (ii) annotate their senses without the need for external sources or manual intervention. We implemented it as a web-based application, which can create large datasets that in turn can be used to train supervised approaches to word sense disambiguation of biomedical acronyms.

AVAILABILITY AND IMPLEMENTATION

The datasets will be generated on demand based on a user query and will be downloadable from https://datainnovation.cardiff.ac.uk/acronyms/.

Biomedical word sense disambiguation with bidirectional long short-term memory and attention-based neural networks

Background

In recent years, deep learning methods have been applied to many natural language processing tasks to achieve state-of-the-art performance. However, in the biomedical domain, they have not out-performed supervised word sense disambiguation (WSD) methods based on support vector machines or random forests, possibly due to inherent similarities of medical word senses.

Results

In this paper, we propose two deep-learning-based models for supervised WSD: a model based on bi-directional long short-term memory (BiLSTM) network, and an attention model based on self-attention architecture. Our result shows that the BiLSTM neural network model with a suitable upper layer structure performs even better than the existing state-of-the-art models on the MSH WSD dataset, while our attention model was 3 or 4 times faster than our BiLSTM model with good accuracy. In addition, we trained “universal” models in order to disambiguate all ambiguous words together. That is, we concatenate the embedding of the target ambiguous word to the max-pooled vector in the universal models, acting as a “hint”. The result shows that our universal BiLSTM neural network model yielded about 90 percent accuracy.

Conclusion

Deep contextual models based on sequential information processing methods are able to capture the relative contextual information from pre-trained input word embeddings, in order to provide state-of-the-art results for supervised biomedical WSD tasks.

Interactive medical word sense disambiguation through informed learning

Objective

Medical word sense disambiguation (WSD) is challenging and often requires significant training with data labeled by domain experts. This work aims to develop an interactive learning algorithm that makes efficient use of expert's domain knowledge in building high-quality medical WSD models with minimal human effort.

Methods

We developed an interactive learning algorithm with expert labeling instances and features. An expert can provide supervision in 3 ways: labeling instances, specifying indicative words of a sense, and highlighting supporting evidence in a labeled instance. The algorithm learns from these labels and iteratively selects the most informative instances to ask for future labels. Our evaluation used 3 WSD corpora: 198 ambiguous terms from Medical Subject Headings (MSH) as MEDLINE indexing terms, 74 ambiguous abbreviations in clinical notes from the University of Minnesota (UMN), and 24 ambiguous abbreviations in clinical notes from Vanderbilt University Hospital (VUH). For each ambiguous term and each learning algorithm, a learning curve that plots the accuracy on the test set against the number of labeled instances was generated. The area under the learning curve was used as the primary evaluation metric.

Results

Our interactive learning algorithm significantly outperformed active learning, the previous fastest learning algorithm for medical WSD. Compared to active learning, it achieved 90% accuracy for the MSH corpus with 42% less labeling effort, 35% less labeling effort for the UMN corpus, and 16% less labeling effort for the VUH corpus.

Conclusions

High-quality WSD models can be efficiently trained with minimal supervision by inviting experts to label informative instances and provide domain knowledge through labeling/highlighting contextual features.

Normalizing acronyms and abbreviations to aid patient understanding of clinical texts: ShARe/CLEF eHealth Challenge 2013, Task 2

BACKGROUND

The ShARe/CLEF eHealth challenge lab aims to stimulate development of natural language processing and information retrieval technologies to aid patients in understanding their clinical reports. In clinical text, acronyms and abbreviations, also referenced as short forms, can be difficult for patients to understand. For one of three shared tasks in 2013 (Task 2), we generated a reference standard of clinical short forms normalized to the Unified Medical Language System. This reference standard can be used to improve patient understanding by linking to web sources with lay descriptions of annotated short forms or by substituting short forms with a more simplified, lay term.

METHODS

In this study, we evaluate 1) accuracy of participating systems' normalizing short forms compared to a majority sense baseline approach, 2) performance of participants' systems for short forms with variable majority sense distributions, and 3) report the accuracy of participating systems' normalizing shared normalized concepts between the test set and the Consumer Health Vocabulary, a vocabulary of lay medical terms.

RESULTS

The best systems submitted by the five participating teams performed with accuracies ranging from 43 to 72 %. A majority sense baseline approach achieved the second best performance. The performance of participating systems for normalizing short forms with two or more senses with low ambiguity (majority sense greater than 80 %) ranged from 52 to 78 % accuracy, with two or more senses with moderate ambiguity (majority sense between 50 and 80 %) ranged from 23 to 57 % accuracy, and with two or more senses with high ambiguity (majority sense less than 50 %) ranged from 2 to 45 % accuracy. With respect to the ShARe test set, 69 % of short form annotations contained common concept unique identifiers with the Consumer Health Vocabulary. For these 2594 possible annotations, the performance of participating systems ranged from 50 to 75 % accuracy.

CONCLUSION

Short form normalization continues to be a challenging problem. Short form normalization systems perform with moderate to reasonable accuracies. The Consumer Health Vocabulary could enrich its knowledge base with missed concept unique identifiers from the ShARe test set to further support patient understanding of unfamiliar medical terms.

Knowledge extraction and semantic annotation of text from the encyclopedia of life

Numerous digitization and ontological initiatives have focused on translating biological knowledge from narrative text to machine-readable formats. In this paper, we describe two workflows for knowledge extraction and semantic annotation of text data objects featured in an online biodiversity aggregator, the Encyclopedia of Life. One workflow tags text with DBpedia URIs based on keywords. Another workflow finds taxon names in text using GNRD for the purpose of building a species association network. Both workflows work well: the annotation workflow has an F1 Score of 0.941 and the association algorithm has an F1 Score of 0.885. Existing text annotators such as Terminizer and DBpedia Spotlight performed well, but require some optimization to be useful in the ecology and evolution domain. Important future work includes scaling up and improving accuracy through the use of distributional semantics.

Applying active learning to supervised word sense disambiguation in MEDLINE

OBJECTIVES

This study was to assess whether active learning strategies can be integrated with supervised word sense disambiguation (WSD) methods, thus reducing the number of annotated samples, while keeping or improving the quality of disambiguation models.

METHODS

We developed support vector machine (SVM) classifiers to disambiguate 197 ambiguous terms and abbreviations in the MSH WSD collection. Three different uncertainty sampling-based active learning algorithms were implemented with the SVM classifiers and were compared with a passive learner (PL) based on random sampling. For each ambiguous term and each learning algorithm, a learning curve that plots the accuracy computed from the test set as a function of the number of annotated samples used in the model was generated. The area under the learning curve (ALC) was used as the primary metric for evaluation.

RESULTS

Our experiments demonstrated that active learners (ALs) significantly outperformed the PL, showing better performance for 177 out of 197 (89.8%) WSD tasks. Further analysis showed that to achieve an average accuracy of 90%, the PL needed 38 annotated samples, while the ALs needed only 24, a 37% reduction in annotation effort. Moreover, we analyzed cases where active learning algorithms did not achieve superior performance and identified three causes: (1) poor models in the early learning stage; (2) easy WSD cases; and (3) difficult WSD cases, which provide useful insight for future improvements.

CONCLUSIONS

This study demonstrated that integrating active learning strategies with supervised WSD methods could effectively reduce annotation cost and improve the disambiguation models.

Applications of natural language processing in biodiversity science

Centuries of biological knowledge are contained in the massive body of scientific literature, written for human-readability but too big for any one person to consume. Large-scale mining of information from the literature is necessary if biology is to transform into a data-driven science. A computer can handle the volume but cannot make sense of the language. This paper reviews and discusses the use of natural language processing (NLP) and machine-learning algorithms to extract information from systematic literature. NLP algorithms have been used for decades, but require special development for application in the biological realm due to the special nature of the language. Many tools exist for biological information extraction (cellular processes, taxonomic names, and morphological characters), but none have been applied life wide and most still require testing and development. Progress has been made in developing algorithms for automated annotation of taxonomic text, identification of taxonomic names in text, and extraction of morphological character information from taxonomic descriptions. This manuscript will briefly discuss the key steps in applying information extraction tools to enhance biodiversity science.

An IR-aided machine learning framework for the BioCreative II.5 Challenge

The team at the University of Wisconsin-Milwaukee developed an information retrieval and machine learning framework. Our framework requires only the standardized training data and depends upon minimal external knowledge resources and minimal parsing. Within the framework, we built our text mining systems and participated for the first time in all three BioCreative II.5 Challenge tasks. The results show that our systems performed among the top five teams for raw F1 scores in all three tasks and came in third place for the homonym ortholog F1 scores for the INT task. The results demonstrated that our IR-based framework is efficient, robust, and potentially scalable.

Linking genes to literature: text mining, information extraction, and retrieval applications for biology

Efficient access to information contained in online scientific literature collections is essential for life science research, playing a crucial role from the initial stage of experiment planning to the final interpretation and communication of the results. The biological literature also constitutes the main information source for manual literature curation used by expert-curated databases. Following the increasing popularity of web-based applications for analyzing biological data, new text-mining and information extraction strategies are being implemented. These systems exploit existing regularities in natural language to extract biologically relevant information from electronic texts automatically. The aim of the BioCreative challenge is to promote the development of such tools and to provide insight into their performance. This review presents a general introduction to the main characteristics and applications of currently available text-mining systems for life sciences in terms of the following: the type of biological information demands being addressed; the level of information granularity of both user queries and results; and the features and methods commonly exploited by these applications. The current trend in biomedical text mining points toward an increasing diversification in terms of application types and techniques, together with integration of domain-specific resources such as ontologies. Additional descriptions of some of the systems discussed here are available on the internet .

Mining experimental evidence of molecular function claims from the literature

MOTIVATION

The rate at which gene-related findings appear in the scientific literature makes it difficult if not impossible for biomedical scientists to keep fully informed and up to date. The importance of these findings argues for the development of automated methods that can find, extract and summarize this information. This article reports on methods for determining the molecular function claims that are being made in a scientific article, specifically those that are backed by experimental evidence.

RESULTS

The most significant result is that for molecular function claims based on direct assays, our methods achieved recall of 70.7% and precision of 65.7%. Furthermore, our methods correctly identified in the text 44.6% of the specific molecular function claims backed up by direct assays, but with a precision of only 0.92%, a disappointing outcome that led to an examination of the different kinds of errors. These results were based on an analysis of 1823 articles from the literature of Saccharomyces cerevisiae (budding yeast).

AVAILABILITY

The annotation files for S.cerevisiae are available from ftp://genome-ftp.stanford.edu/pub/yeast/data_download/literature_curation/gene_association.sgd.gz. The draft protocol vocabulary is available by request from the first author.