From published academic research papers to patents filed in international patent databases, there are large amounts of
publicly available data from open literature sources that are not currently being used. Northstar, a new company based in
the Liverpool City Region, wanted to find a way to capture relevant data to repurposes scientific inventions, generate new
ideas and patents more quickly and cost-effectively without relying on human researchers and information scientists. They
wanted to understand how they could use AI to create a digital platform capable of acting as a virtual research scientist.
Funded by ERDF as part of the LCR4 START programme, Northstar worked with data scientists and AI experts at the
Science and Technology Facilities Council (STFC) Hartree Centre to create a workflow capable of bridging a gap between
the science and the application. The team had expertise in various aspects of data science, from data engineering and data
standardisation to data exploitation, using their skills to deliver a proof-of-concept AI demonstrator.
This tool is capable of searching the open literature in a fraction of the time and cost of a manual approach. Using AI
to dive through the breadth and depth of scientific literature rather than humans selectively searching a narrow
field of patents and papers offers greater breadth and depth to scientific discovery. This has the potential to
significantly accelerate the pathway to scientific discovery for companies who are routinely looking for
new inventions by removing the limitations of what traditional desk-based research can highlight.
Beyond LCR4 START, Northstar plans to continue this project. Now the company has a live
demonstrator, it is working on creating a user interface and web portal for the tool that
embeds machine learning to self-learn from both questions and answers, outputting
results in accessible and graphical formats for researchers.
“All great innovations are built on what has happened
previously. We wanted to find a way to capture the
relevant open data that would enable us to
repurpose scientific inventions quicker
and generate new ideas.This work can speed up the process of creating new
patents and methodologies that can make the world
a better place. Our ultimate goal is to bring the dark
scientific data into the light and demonstrate the
first AI at scale to achieve this.”