AL-MedLit: an active learning collaborative platform for biomedical literature classification and model evolution

Ohio State University, Columbus OH

Investigators

Abstract

Project Summary The goal of this application is to support Lei Wang, Ph.D., as a Research Scientist and Research Software Specialist at The Ohio State University. He will collaborate with Dr. Lang Li and develop an active learning collaborative platform for language models to support several NIH-funded research programs (R01LM014199, P30HD106451, and U01CA248240). The proposed AL-MedLit platform is the first software that allows language models and human experts to collaborate, ensuring the improvement of knowledgebase development and reduction of the curation cost. The innovative algorithms, including both conventional uncertainty sampling schemes and novel sampling schemes such as positive sampling, negative sampling, and sample calibration, will be implemented in the platform to facilitate the evolution of language models. This Research Software Engineer award would allow 3 years of funding to fully maintain Dr. Lei Wang’s dedicated efforts toward developing an innovative active learning platform for pharmaco-informatics research programs. In particular, he will expand his skill set, especially in the rapidly evolving fields of natural language processing and artificial intelligence and its applications in pharmaco-informatics research. Results will be broadly disseminated through local and national meetings as well as peer-reviewed journal manuscripts.