Bio-inspired techniques for pattern recognition

Intelligence as it occurs in nature demands a level of energy efficiency, speed, accuracy, and generalisability that current algorithms are not readily capable of emulating independently. There are also application areas such as swarm robotics or wireless sensor networks that rely upon relatively simple computational units to carry out complex tasks. Such applications operate under similar constraints (energy efficiency, speed, accuracy and generalisability) to biological organisms and can thus benefit from the computational strategies employed by nature. This thesis addresses the problems outlined above by focusing on information from neuroscience and existing bio-inspired approaches.