Can we make the self-driving cars even smarter? Can we make them act appropriately even on unlikely events? Sara Mahmoud, PhD student in Informatics at the University of Skövde, has investigated how cognitively inspired design can be used to create smarter AI systems in cars.
Sara Mahmoud has investigated how artificial systems can be created so that they gain cognitive abilities and can perform intelligent tasks.
A self-driving car must know all traffic rules. It needs to understand what it sees and make quick decisions. That in itself is a very complex process. And sometimes unexpected things also happen in traffic. However, unexpected events are something that we humans are good at handling. We can use our intuition and past experiences, we have emotional intelligence and can think outside the box. Here, the AI systems have something to learn! Is it possible to teach an AI to think more like a human? Sara Mahmoud has taken a closer look at this in her research.
Sara has investigated how artificial systems can be created so that they gain cognitive abilities and can perform intelligent tasks, by being inspired by human intelligence. The use of cognitive theories within applied systems is called cognitively inspired design, CID.
“My study examines how we can transform cognitive theories so that they can be used when various smart systems are to be designed, how the theories can be used in the real world. I have chosen self-driving cars as the main application area,” says Sara Mahmoud.
She makes cognitive inspiration concrete
Her research findings highlight opportunities and challenges of cognitively inspired design as well as limitations of the current AI methods used in self-driving cars.
With the research, she has created a framework that breaks down the somewhat abstract concept of cognitive inspiration. It can be seen as a guide to explore and apply ideas about human thinking and cognition in the creation of various products and systems.
“Instead of dealing with cognitive inspiration as a big black box, the researcher or developer using my framework can understand how to use the knowledge when designing their system or product.”
The framework for cognitively inspired design thus establishes guidelines for using cognitive inspiration for system design. Unlike the views that either regard cognitive theories as the definitive solution for super intelligence or that consider that AI should be isolated from the field of cognition, Sara's research takes a pragmatic view that cognitive science can provide new inspiration for system design, while bringing the challenges and the difficulties in the process.
What's next for you?
“Since my research is about inspiration info system design, I would like to gain more industrial and research experiences in the field of engineering, and I have started a job in the automotive industry. I would like to apply what I have learned in cognitive systems and engineering approaches in that field. I have always had an ambition to combine research and industry and I aim to continue with that.
Sara Mahmoud defends her thesis "Cognitively inspired design: Re-think the wheel for self-driving cars" on Monday 18 December at the University of Skövde.