Mark2Cure

Mark2Cure leverages citizen science to perform named entity recognition (NER), apply standardized vocabularies for normalization, and extract relationships from biomedical literature to support literature mining and accelerate biomedical discovery.

Key Features:

• Named Entity Recognition (NER): Performs named entity recognition on free-text biomedical literature to identify biomedical entities.
• Normalization: Maps identified entities to standardized vocabularies to ensure consistent term usage.
• Relationship Extraction: Identifies relationships between entities within text to support construction of entity associations.
• Evaluation Metrics: Uses the Zooniverse Matrices of Citizen Science Success to evaluate annotation quality and project effectiveness.

Scientific Applications:

• Biomedical literature mining: Enables extraction of entities and relationships from the biomedical literature to organize knowledge for research.
• Accelerating discovery: Supports faster interpretation of literature to potentially accelerate development of cures for diseases.
• Citizen science NLP model: Serves as a model for natural language processing (NLP) tasks within citizen science frameworks by leveraging volunteer-contributed annotations.

Methodology:

Employs a citizen science crowdsourcing model to perform named entity recognition, references prior crowdsourcing demonstrations using Amazon Mechanical Turk (AMT), and applies the Zooniverse Matrices of Citizen Science Success for evaluation.