Curiosity and play to enhance AI’s learning ability 

Guided by curiosity, play and empathy, Wallenberg Scholar Amy Loutfi and her research team aim to study and develop the learning process in embodied AI systems, such as robots.

Conventional artificial intelligence (AI) processes and makes decisions based on large data sets. The concept of embodied AI entails providing an AI system with a physical form, like a robot, drone, or self-driving car, allowing it to interact with and learn from its environment.

“By interacting with their surroundings, embodied AI systems can collect information about the world around it and use this information to adapt and enhance its performance,” states Amy Loutfi, professor of computer science at Örebro University.

Curiosity, play and empathy inspires research

Amy Loutfi and her team are dedicated to developing new techniques and algorithms that enhance the learning process in embodied AI. They will draw upon three specific learning concepts within embodied AI to inspire their research: curiosity, play and empathy.

“Play is a natural and essential activity to facilitate learning and the acquisition of new skills. Drawing on this inspiration, we aim to explore how language and playfulness can enhance the learning process. Allowing AI to envision new and abstract goals through language enables an AI to explore the world and even design its own chosen path of study or target sequence,” explains Amy Loutfi.

Artificial intelligence is developing at an accelerating pace.

“This is why emphasizing basic research into intelligence and learning is central to my goal of taking the next steps in artificial intelligence,” says Amy Loutfi.