Stockholm university

AI finds the best ideas

Ideas are available in abundance, but how do you know which ones to exploit? Let AI and machine learning lead the way. Workneh Yilma Ayele explains how it can be done in his doctoral thesis.

Workneh Yilma Ayele. Photo: Åse Karlén.
Workneh Yilma Ayele. Photo: Åse Karlén.

Ideas that find their audience can grow into innovations. But it’s not easy to know which ones will be successful – or when success will happen. Leonardo da Vinci designed flying machines in the 15th century, but his ideas were not realized until the combustion engine was invented.

Nikola Tesla (1856–1943) is another inventor who was ahead of his time. His ideas on free energy and global, wireless communication were considered madness. Nowadays, they are commodities that we take for granted.

A toolbox for ideas

Both da Vinci and Tesla are mentioned in Workneh Yilma Ayele’s doctoral thesis that presents a toolbox for idea generation and evaluation. The tools use machine learning and AI, and they build upon data driven and contest driven processes.

“Today, large amounts of data are generated. They include data from news articles, patents, research reports and discussions in social media. Such text-based data provides input to my machine learning models”, says Workneh Yilma Ayele who is the first person to present his thesis at the Department of Computer and Systems Sciences, Stockholm University, in 2022.

Manual systems for handling ideas have been around for a long time, but today’s technology delivers more sophisticated solutions. AI can work with huge data sets and discover patterns. This sheds light on interesting ideas, but the artificial intelligence needs help from human intelligence – and intuition.

“Data driven idea generation is a very effective and productive process. It can provide us with ideas and innovations that will make our lives better, if we use human-centered AI”, says Workneh Yilma Ayele.

The ability to continuously generate good ideas is an important factor for competing companies.

“Large companies like Starbucks and Lego have their own systems for collecting ideas from emloyees and customers. Lakefront Brewery in the US chose to analyze comments on the social platform Reddit. That led to the decision to launch a gluten free beer. Innovation competitions and hackathons are other common ways to gather ideas.”

Sweden excels in innovation

Workneh Yilma Ayele emphasizes the importance of established companies working together with ambitious startups. Outsourcing parts of their R&D to startups could be a way to stay in the lead. This is something that Kodak, market leading at the time, failed to do when digitalization disrupted the photo industry.

“Sweden is always in the top three when innovation abilities are ranked globally. I think we need to do better in commercialization, turning ideas into real value.”

“Finding the best ideas is never easy, but everyone has a need for good ideas”, says Workneh Yilma Ayele.

About the research

Workneh Yilma Ayele presented his thesis on March 24, 2022. Gustaf Juell-Skielse is the supervisor, Paul Johannesson is the co-supervisor.

The thesis is published online: “A toolbox for idea generation and evaluation. Machine learning, data-driven, and contest-driven approaches to support idea generation”

Workneh Yilma Ayele continues to work as a researcher and teacher at the Department of Computer and Systems Sciences (DSV).
Read about Workneh Yilma Ayele’s research

In the latest ranking from WIPO, the global forum for intellectual property of the United Nations, Sweden reached second place.
Read about the Global Innovation Index (GII) on WIPO’s website

About the Department of Computer and Systems Sciences (DSV)

Text: Åse Karlén