My research interests are in neural algorithms, whether biological or artificial. I tend to focus on how biological neural circuits might implement or approximate mathematical algorithms used in machine learning. I also take principles of neural function and use them to augment artificial neural networks for machine learning applications.
I often use insect models of learning, decision-making, and behaviour, with a focus on the mushroom body (a key region for learning in insects), partly because of the powerful genetic tools that are available for experiments in Drosophila (fruit flies), and partly because of their fascinating adaptations and importance for biodiversity.
My recent work has addressed reinforcement learning models of the insect mushroom body, as well as representation learning in insect-like neural architectures.