Skip to main content

The artificial intelligence of IRIS

By January 3, 2019August 5th, 2021sdg 17, sdg 9, technology

Anita Schjøll Brede explains how artificial intelligence like Iris could simplify scientific research and much more.

CRISTINA: Anita can you give us a metaphor to understand exactly what AI is and how does it work?

ANITA: We’re shaping these systems more like a human brain and less like a calculator, so say that you’re running a pizzeria and every morning you have to figure out how much dough you should make for all the pizzas today. It’s a complex question, what is the weather? What historically have I sold? Is it tourist season or not? What day of the week is it? There’s a number of data points that are important for that and you know with human experience you can make a fair assessment, but if you have a system that learns over time you can set it up to calculate for you every morning, based on historical data, how much you should make. And every day you have new data to feed the system and the system will learn every single day and become better and better at predicting it.

CRISTINA: Now how are you applying AI?

ANITA: We have a variety of tools that help researchers go from: I have this problem I need to solve, map out all existing literature, and narrow that down to a very precise reading list.


ANITA: And this is incredibly time consuming. We have 150 million research papers, millions and millions of patents in the world and one human being simply cannot read and understand everything. So what we’re doing is building a system that allows us to actually do that, where one human can sit down and find exactly the pieces of the puzzle that they need to solve the problem.

CRISTINA: Do you think that will spare time in maybe doing research on things that have already been researched and conclusions that have made sense been drawn?

ANITA: Exactly and so much of the solutions to our really big problems, whether we talk about climate change or any of the big problems we’re facing, it’s interdisciplinary. It’s taking that solution and that solution and that one and piecing it all together and that is not possible for a human being to do today, but we’re making it possible for an AI to do so.

CRISTINA: And it’s functioning already?

ANITA: Part of it. The part of the tool that is helping you do the literature reviews and going from problem to precise reading list. That is functional, we’re selling it to universities and to big corporates as well. And then the next step is to start extracting a hypothesis and finding conclusions and starting to see the big patterns.

CRISTINA: So it will become a researcher itself?

ANITA: Eventually yes.