3 December 2019 in Journal of the American Medical Informatics Association
Objective: This article methodically reviews the literature on deep learning (DL) for natural language processing (NLP) in the clinical domain, providing quantitative analysis to answer 3 research questions concerning methods, scope, and context of current research.Materials and Methods: We searched MEDLINE, EMBASE, Scopus, the Association for Computing Machinery Digital Library, and the Association for Computational Linguistics Anthology for articles using DL-based approaches to NLP problems in electronic health records. After screening 1,737 articles, we collected data on 25 variables across 212 papers.Results: DL in clinical NLP publications more than doubled each year, through 2018. Recurrent neural networks (60.8%) and word2vec embeddings (74.1%) were the most popular methods; the information extraction tasks of text classification, named entity recognition, and relation extraction were dominant (89.2%). However, there was a “long tail” of other methods and specific tasks. Most contributions were methodological variants or applications, but 20.8% were new methods of some kind. The earliest adopters were in the NLP community, but the medical informatics community was the most prolific.Discussion: Our analysis shows growing acceptance of deep learning as a baseline for NLP research, and of DL-based NLP in the medical community. A number of common associations were substantiated (eg, the preference of recurrent neural networks for sequence-labeling named entity recognition), while others were surprisingly nuanced (eg, the scarcity of French language clinical NLP with deep learning).Conclusion: Deep learning has not yet fully penetrated clinical NLP and is growing rapidly. This review highlighted both the popular and unique trends in this active field.
2 July 2019 in Journal of the American Medical Informatics Association
Objective: Neural network–based representations (“embeddings”) have dramatically advanced natural language processing (NLP) tasks, including clinical NLP tasks such as concept extraction. Recently, however, more advanced embedding methods and representations (eg, ELMo, BERT) have further pushed the state of the art in NLP, yet there are no common best practices for how to integrate these representations into clinical tasks. The purpose of this study, then, is to explore the space of possible options in utilizing these new models for clinical concept extraction, including comparing these to traditional word embedding methods (word2vec, GloVe, fastText).Materials and Methods: Both off-the-shelf, open-domain embeddings and pretrained clinical embeddings from MIMIC-III (Medical Information Mart for Intensive Care III) are evaluated. We explore a battery of embedding methods consisting of traditional word embeddings and contextual embeddings and compare these on 4 concept extraction corpora: i2b2 2010, i2b2 2012, SemEval 2014, and SemEval 2015. We also analyze the impact of the pretraining time of a large language model like ELMo or BERT on the extraction performance. Last, we present an intuitive way to understand the semantic information encoded by contextual embeddings.Results: Contextual embeddings pretrained on a large clinical corpus achieves new state-of-the-art performances across all concept extraction tasks. The best-performing model outperforms all state-of-the-art methods with respective F1-measures of 90.25, 93.18 (partial), 80.74, and 81.65.Conclusions: We demonstrate the potential of contextual embeddings through the state-of-the-art performance these methods achieve on clinical concept extraction. Additionally, we demonstrate that contextual embeddings encode valuable semantic information not accounted for in traditional word representations.
1 March 2021 in Journal of Biomedical Informatics
8 July 2020 in Journal of the American Medical Informatics Association
TREC-COVID is an information retrieval (IR) shared task initiated to support clinicians and clinical research during the COVID-19 pandemic. IR for pandemics breaks many normal assumptions, which can be seen by examining 9 important basic IR research questions related to pandemic situations. TREC-COVID differs from traditional IR shared task evaluations with special considerations for the expected users, IR modality considerations, topic development, participant requirements, assessment process, relevance criteria, evaluation metrics, iteration process, projected timeline, and the implications of data use as a post-task test collection. This article describes how all these were addressed for the particular requirements of developing IR systems under a pandemic situation. Finally, initial participation numbers are also provided, which demonstrate the tremendous interest the IR community has in this effort.
1 January 2020 in Database: The Journal of Biological Databases and Curation
It is a growing trend among researchers to make their data publicly available for experimental reproducibility and data reusability. Sharing data with fellow researchers helps in increasing the visibility of the work. On the other hand, there are researchers who are inhibited by the lack of data resources. To overcome this challenge, many repositories and knowledge bases have been established to date to ease data sharing. Further, in the past two decades, there has been an exponential increase in the number of datasets added to these dataset repositories. However, most of these repositories are domain-specific, and none of them can recommend datasets to researchers/users. Naturally, it is challenging for a researcher to keep track of all the relevant repositories for potential use. Thus, a dataset recommender system that recommends datasets to a researcher based on previous publications can enhance their productivity and expedite further research. This work adopts an information retrieval (IR) paradigm for dataset recommendation. We hypothesize that two fundamental differences exist between dataset recommendation and PubMed-style biomedical IR beyond the corpus. First, instead of keywords, the query is the researcher, embodied by his or her publications. Second, to filter the relevant datasets from non-relevant ones, researchers are better represented by a set of interests, as opposed to the entire body of their research. This second approach is implemented using a non-parametric clustering technique. These clusters are used to recommend datasets for each researcher using the cosine similarity between the vector representations of publication clusters and datasets. The maximum normalized discounted cumulative gain at 10 (NDCG@10), precision at 10 (p@10) partial and p@10 strict of 0.89, 0.78 and 0.61, respectively, were obtained using the proposed method after manual evaluation by five researchers. As per the best of our knowledge, this is the first study of its kind on content-based dataset recommendation. We hope that this system will further promote data sharing, offset the researchers’ workload in identifying the right dataset and increase the reusability of biomedical datasets. Database URL: http://genestudy.org/recommends/#/
1 December 2019 in Journal of Biomedical Informatics