Virk: An active learning-based system for bootstrapping knowledge base development in the neurosciences

Kyle H. Ambert, Aaron Cohen, Gully A P C Burns, Eilis Boudreau, Mustafa (Kemal) Sonmez

Research output: Contribution to journalArticle

3 Citations (Scopus)

Abstract

The frequency and volume of newly-published scientific literature is quickly making manual maintenance of publicly-available databases of primary data unrealistic and costly. Although machine learning (ML) can be useful for developing automated approaches to identifying scientific publications containing relevant information for a database, developing such tools necessitates manually annotating an unrealistic number of documents. One approach to this problem, active learning (AL), builds classification models by iteratively identifying documents that provide the most information to a classifier. Although this approach has been shown to be effective for related problems, in the context of scientific databases curation, it falls short. We present Virk, an AL system that, while being trained, simultaneously learns a classification model and identifies documents having information of interest for a knowledge base. Our approach uses a support vector machine (SVM) classifier with input features derived from neuroscience-related publications from the primary literature. Using our approach, we were able to increase the size of the Neuron Registry, a knowledge base of neuron-related information, by a factor of 90%, a knowledge base of neuron-related information, in 3 months. Using standard biocuration methods, it would have taken between 1 and 2 years to make the same number of contributions to the Neuron Registry. Here, we describe the system pipeline in detail, and evaluate its performance against other approaches to sampling in AL.

Original languageEnglish (US)
Article number38
JournalFrontiers in Neuroinformatics
Volume7
Issue numberDEC
DOIs
StatePublished - Dec 25 2013

Fingerprint

Problem-Based Learning
Knowledge Bases
Neurosciences
Neurons
Databases
Registries
Learning systems
Classifiers
Literature
Support vector machines
Publications
Pipelines
Maintenance
Sampling

Keywords

  • Active learning
  • Biocuration
  • Community-curated database
  • Machine learning
  • Neuroinformatics
  • Text-mining

ASJC Scopus subject areas

  • Neuroscience (miscellaneous)
  • Biomedical Engineering
  • Computer Science Applications

Cite this

Virk : An active learning-based system for bootstrapping knowledge base development in the neurosciences. / Ambert, Kyle H.; Cohen, Aaron; Burns, Gully A P C; Boudreau, Eilis; Sonmez, Mustafa (Kemal).

In: Frontiers in Neuroinformatics, Vol. 7, No. DEC, 38, 25.12.2013.

Research output: Contribution to journalArticle

@article{b529dfac91ef485d8ee0c71dd3e8f577,
title = "Virk: An active learning-based system for bootstrapping knowledge base development in the neurosciences",
abstract = "The frequency and volume of newly-published scientific literature is quickly making manual maintenance of publicly-available databases of primary data unrealistic and costly. Although machine learning (ML) can be useful for developing automated approaches to identifying scientific publications containing relevant information for a database, developing such tools necessitates manually annotating an unrealistic number of documents. One approach to this problem, active learning (AL), builds classification models by iteratively identifying documents that provide the most information to a classifier. Although this approach has been shown to be effective for related problems, in the context of scientific databases curation, it falls short. We present Virk, an AL system that, while being trained, simultaneously learns a classification model and identifies documents having information of interest for a knowledge base. Our approach uses a support vector machine (SVM) classifier with input features derived from neuroscience-related publications from the primary literature. Using our approach, we were able to increase the size of the Neuron Registry, a knowledge base of neuron-related information, by a factor of 90{\%}, a knowledge base of neuron-related information, in 3 months. Using standard biocuration methods, it would have taken between 1 and 2 years to make the same number of contributions to the Neuron Registry. Here, we describe the system pipeline in detail, and evaluate its performance against other approaches to sampling in AL.",
keywords = "Active learning, Biocuration, Community-curated database, Machine learning, Neuroinformatics, Text-mining",
author = "Ambert, {Kyle H.} and Aaron Cohen and Burns, {Gully A P C} and Eilis Boudreau and Sonmez, {Mustafa (Kemal)}",
year = "2013",
month = "12",
day = "25",
doi = "10.3389/fninf.2013.00038",
language = "English (US)",
volume = "7",
journal = "Frontiers in Neuroinformatics",
issn = "1662-5196",
publisher = "Frontiers Research Foundation",
number = "DEC",

}

TY - JOUR

T1 - Virk

T2 - An active learning-based system for bootstrapping knowledge base development in the neurosciences

AU - Ambert, Kyle H.

AU - Cohen, Aaron

AU - Burns, Gully A P C

AU - Boudreau, Eilis

AU - Sonmez, Mustafa (Kemal)

PY - 2013/12/25

Y1 - 2013/12/25

N2 - The frequency and volume of newly-published scientific literature is quickly making manual maintenance of publicly-available databases of primary data unrealistic and costly. Although machine learning (ML) can be useful for developing automated approaches to identifying scientific publications containing relevant information for a database, developing such tools necessitates manually annotating an unrealistic number of documents. One approach to this problem, active learning (AL), builds classification models by iteratively identifying documents that provide the most information to a classifier. Although this approach has been shown to be effective for related problems, in the context of scientific databases curation, it falls short. We present Virk, an AL system that, while being trained, simultaneously learns a classification model and identifies documents having information of interest for a knowledge base. Our approach uses a support vector machine (SVM) classifier with input features derived from neuroscience-related publications from the primary literature. Using our approach, we were able to increase the size of the Neuron Registry, a knowledge base of neuron-related information, by a factor of 90%, a knowledge base of neuron-related information, in 3 months. Using standard biocuration methods, it would have taken between 1 and 2 years to make the same number of contributions to the Neuron Registry. Here, we describe the system pipeline in detail, and evaluate its performance against other approaches to sampling in AL.

AB - The frequency and volume of newly-published scientific literature is quickly making manual maintenance of publicly-available databases of primary data unrealistic and costly. Although machine learning (ML) can be useful for developing automated approaches to identifying scientific publications containing relevant information for a database, developing such tools necessitates manually annotating an unrealistic number of documents. One approach to this problem, active learning (AL), builds classification models by iteratively identifying documents that provide the most information to a classifier. Although this approach has been shown to be effective for related problems, in the context of scientific databases curation, it falls short. We present Virk, an AL system that, while being trained, simultaneously learns a classification model and identifies documents having information of interest for a knowledge base. Our approach uses a support vector machine (SVM) classifier with input features derived from neuroscience-related publications from the primary literature. Using our approach, we were able to increase the size of the Neuron Registry, a knowledge base of neuron-related information, by a factor of 90%, a knowledge base of neuron-related information, in 3 months. Using standard biocuration methods, it would have taken between 1 and 2 years to make the same number of contributions to the Neuron Registry. Here, we describe the system pipeline in detail, and evaluate its performance against other approaches to sampling in AL.

KW - Active learning

KW - Biocuration

KW - Community-curated database

KW - Machine learning

KW - Neuroinformatics

KW - Text-mining

UR - http://www.scopus.com/inward/record.url?scp=84891541647&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84891541647&partnerID=8YFLogxK

U2 - 10.3389/fninf.2013.00038

DO - 10.3389/fninf.2013.00038

M3 - Article

AN - SCOPUS:84891541647

VL - 7

JO - Frontiers in Neuroinformatics

JF - Frontiers in Neuroinformatics

SN - 1662-5196

IS - DEC

M1 - 38

ER -