Bio-inspired AI

Bio Robot

Animals do not do Deep Learning. Yet, they have evolved to solve challenging problems using simple and robust algorithms. It's those algorithms we are after.” Thomas Nowotny
Professor of Informatics

How does adaptive behaviour emerge from natural systems? How can we build intelligent systems? In the AI research group we use computational tools and methods to analyse biological systems and explore their capacity for problem solving in natural environments. We also employ biologically-inspired engineering to replicate animals’ capabilities in autonomous systems to both, build more robust robots and shed light on the biological basis of intelligence. By embedding Neural Models derived from natural systems, onto hardware devices, we aim to understand how artificial systems can sense, learn and move through complex natural environments. In separate strands of work we use evolution-inspired algorithms for optimisation, in particular to evolve robot controllers - evolutionary robotics - or, to evolve electronic hardware. In both cases, suprising, new solutions emerge. Further topics are artificial life and large scale neural network models for Computational Neuroscience, bio-mimetic AI and neuromorphic computing. Read on below to find out more. 

Bio-inspired robot control


Artificial Life and Adaptive Systems


Evolutionary Robotics and Evolutionary Hardware


Spiking neural networks and Neuromorphic Computing