Text mining in the Cloud

点击翻译按钮获取中文摘要

Project Summary The Alliance of Genome Resources (Alliance) is developing shared, sustainable infrastructure for the curation, storage, analysis, and presentation of genomic and genetic data about research organisms to serve biomedical researchers, bioinformaticians, and artificial intelligence (AI) and machine learning (ML) researchers, as well as clinicians, students and teachers. We propose to develop a unified cloud-based text-mining service to enable AI/ML approaches to identify relevant documents and text-spans suitable for use by professional biocurators, authors who curate their own papers pre- or post-publication, and researchers who want sentence-level full text search. This project will leverage the PubMedCentral (PMC) cloud-based open-access corpus by implementing a set of neural network classification and NLP algorithms in the cloud to take advantage of the PubMed Central Open Access (PMC-OA) corpus already in the cloud. The project will implement in the cloud software developed by the Textpresso group of the Alliance, and carry out computationally intensive indexing of papers using neural networks to classify papers, Alliance-custom entity recognition, and Textpresso ontology-based indexing to aid biocuration. The project will be sustained by the Alliance and the MODs.