ITEA is the Eureka Cluster on software innovation
ITEA is the Eureka Cluster on software innovation


Fully Automated AI Data Extraction from Scientific Literature

Project description

We intend to compare the performance and accuracy of applying up to four AI LLMs (large language models) and a combination of AI deterministic and LLM methods to fully extract text from scientific publications. The project seeks to establish methodological standards for AI LLM's use to reduce the time it takes to complete a systematic literature review, accelerate the time-to-market impact of SLRs’ use for regulatory submissions, and lower reviewer fatigue and burnout from their overall manual conduct of SLRs. We will achieve these goals in the following way: 1. Apply AI methodologies and techniques to improve the speed, accuracy, and efficiency of scientific literature reviews, which take on average 69 weeks to complete manually and are often error-prone 2. Leverage workflow automation and integrated AI technologies to autonomously train AI/ML models, including no-code AI/ML 3. Deploy selected AI technologies to extract relevant data from scientific literature at scale and speed 4, Enable non-AI experts to train and deploy highly targeted AI to autonomously extract data more accurately from large volumes of scientific literature

Project leader

Chris Stephen Wright
DistillerSR Inc., Canada
Alt Alt Alt Alt Alt

Project publications