Iris.ai - extracting the essence of research

Offentlig prosjektsammendrag

We human beings have access to an unprecedented amount of scientific knowledge; 6000 papers and 8000 patents are published every single day. This is both good news and bad news. Good because it pushes humanity forward; eradicates diseases; solves climate challenges - bad, because it is becoming increasingly overwhelming for even the smartest researchers and innovators among us to try and navigate it all. The Iris.ai team has already launched the first version of the Researcher Workspace - a tool suite to radically speed up researchers' handling of all this information. With this research project, we aim to take the next step beyond the current frontier. We are now building a system that can efficiently retrieve and combine knowledge of various forms from different research domains. This knowledge can exist in textual forms, such as written texts in scientific publications, in non-textual forms, such as tabular data, which indicate causal relationships between entities. Our research goal until the end of 2022 is to understand how to systematically represent all this knowledge, including from tables and in the shape of scientific summaries. This means we can expand the Researcher Workspace from a great time savings tool for finding the right literature, to also distilling the key elements of each vital document at the click of a finger. We are operating on the forefront of the field of Natural Language Processing, bringing powerful tools to the hand of some of our most capable and important contributors to progress - the scientists and researchers.