Extracting Evidence Fragments for Distant Supervision of Molecular Interactions

Burns, Gully A. P. C.; Dasigi, Pradeep; Hovy, Eduard

doi:10.1101/192856

Cited by 4 publications

(4 citation statements)

References 15 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…We sought to characterize the performance of several different standard neural network approaches in order to support the use of these tools across various similar focused text classification tasks of relevance to biocuration. This directly builds on previous work performing the same classification task with non-deep learning methods (21) or with FastText classifiers as part of a broader workflow (18). We here examine classification performance for all available word embedding models under a variety of parameterizations (see Figure 2).…”

Section: Resultsmentioning

confidence: 99%

“…We queried PubMed to obtain title, abstract and MeSH data for each paper. We downloaded * .nxml-formatted files from PMC and invoked the UIMA-Bioc preprocessing pipeline (as described previously) to extract figure captions and ‘evidence fragment’ text pertaining to each subfigure mentioned in the textual narrative (21). Given these data, we performed document-level text classification as a ‘triage’ task (i.e.…”

Section: Methodsmentioning

confidence: 99%

“…We investigated four cases (i) using each paragraph in the full-text document, (ii) using each ‘evidence fragment’ (i.e. the text surrounding a figure reference that directly describes the figure) (21), (iii) using each figure caption and (iv) using evidence fragments and captions together. We broke the document-level text into collections of paragraphs so that the network classifier could make inferences aggregated over classification decisions made for each individual paragraph.…”

Section: Methodsmentioning

confidence: 99%

See 2 more Smart Citations

Building deep learning models for evidence classification from the open access biomedical literature

Burns

Peng

2019

Database

Self Cite

View full text Add to dashboard Cite

We investigate the application of deep learning to biocuration tasks that involve classification of text associated with biomedical evidence in primary research articles. We developed a large-scale corpus of molecular papers derived from PubMed and PubMed Central open access records and used it to train deep learning word embeddings under the GloVe, FastText and ELMo algorithms. We applied those models to a distant supervised method classification task based on text from figure captions or fragments surrounding references to figures in the main text using a variety or models and parameterizations. We then developed document classification (triage) methods for molecular interaction papers by using deep learning mechanisms of attention to aggregate classification-based decisions over selected paragraphs in the document. We were able to obtain triage performance with an accuracy of 0.82 using a combined convolutional neural network, bi-directional long short-term memory architecture augmented by attention to produce a single decision for triage. In this work, we hope to encourage biocuration systems developers to apply deep learning methods to their specialized tasks by repurposing large-scale word embedding to apply to their data.

show abstract

Section: Resultsmentioning

confidence: 99%

Section: Methodsmentioning

confidence: 99%

Section: Methodsmentioning

confidence: 99%

See 1 more Smart Citation

Building deep learning models for evidence classification from the open access biomedical literature

Burns

Peng

2019

Database

Self Cite

View full text Add to dashboard Cite

show abstract

“…Integrating computational approaches into the workflow of biocuration can be seen in many applications such as constructing genomics knowledge base (Baumgartner Jr et al 2007), biomedical document classification (Cohen 2006;Shatkay, Chen, and Blostein 2006;Jiang et al 2017;Simon et al 2019), biomedical text mining (Dowell et al 2009), and human-in-the-loop curation (Lee et al 2018). Some prior works also adopt multimodal machine learning for general biomedical information extractions (Schlegl et al 2015;Eickhoff et al 2017;Zhang et al 2017), as well as textual extraction (Burns, Dasigi, and Hovy 2017), medical image captioning (Shin et al 2016), and automated diagnosis from medical images (Jing, Xie, and Xing 2018;Wang et al 2018;Liu et al 2019a).…”

Section: Related Workmentioning

confidence: 99%

MELINDA: A Multimodal Dataset for Biomedical Experiment Method Classification

Singh

Paul

et al. 2021

AAAI

View full text Add to dashboard Cite

We introduce a new dataset, MELINDA, for Multimodal biomEdicaL experImeNt methoD clAssification. The dataset is collected in a fully automated distant supervision manner, where the labels are obtained from an existing curated database, and the actual contents are extracted from papers associated with each of the records in the database. We benchmark various state-of-the-art NLP and computer vision models, including unimodal models which only take either caption texts or images as inputs, and multimodal models. Extensive experiments and analysis show that multimodal models, despite outperforming unimodal ones, still need improvements especially on a less-supervised way of grounding visual concepts with languages, and better transferability to low resource domains. We release our dataset and the benchmarks to facilitate future research in multimodal learning, especially to motivate targeted improvements for applications in scientific domains.

show abstract

Scientific Discourse Tagging for Evidence Extraction

Li¹,

Burns²,

Peng³

2021

Proceedings of the 16th Conference of the European Chapter of the Association for Computational Linguistics: Main Volume

View full text Add to dashboard Cite

Evidence plays a crucial role in any biomedical research narrative, providing justification for some claims and refutation for others. We seek to build models of scientific argument using information extraction methods from fulltext papers. We present the capability of automatically extracting text fragments from primary research papers that describe the evidence presented in that paper's figures, which arguably provides the raw material of any scientific argument made within the paper. We apply richly contextualized deep representation learning pre-trained on biomedical domain corpus to the analysis of scientific discourse structures and the extraction of "evidence fragments" (i.e., the text in the results section describing data presented in a specified subfigure) from a set of biomedical experimental research articles. We first demonstrate our state-of-the-art scientific discourse tagger on two scientific discourse tagging datasets and its transferability to new datasets. We then show the benefit of leveraging scientific discourse tags for downstream tasks such as claim-extraction and evidence fragment detection. Our work demonstrates the potential of using evidence fragments derived from figure spans for improving the quality of scientific claims by cataloging, indexing and reusing evidence fragments as independent documents.

show abstract

Extracting Evidence Fragments for Distant Supervision of Molecular Interactions

Cited by 4 publications

References 15 publications

Building deep learning models for evidence classification from the open access biomedical literature

Building deep learning models for evidence classification from the open access biomedical literature

MELINDA: A Multimodal Dataset for Biomedical Experiment Method Classification

Scientific Discourse Tagging for Evidence Extraction

Contact Info

Product

Resources

About