Hello! I am a postdoctoral researcher working on (artificial) intelligence in the Brain & Cognitive Sciences department of MIT. My principal interests lie in bridging the gap between the performance of current deep learning systems and the abilities and behaviours shown by biological systems, in particular in improving the ability to learn continually and rapidly from small amounts of data. Currently I am working mostly on understanding the behaviour of neural networks with a little bit of work on model-based reinforcement learning as well. I work primarily with Xavier Boix in the Sinha Lab for Developmental Research.
I recently submitted my PhD thesis, working in the Institute of Perception, Action and Behaviour at the University of Edinburgh where I was supervised by Dr. Taku Komura. During my studies I first developed methods in deep learning applications with a focus on data efficient modelling of style and content in 3D humanoid animation data, that is rapidly adapting to new domains in time-series data. More recently I focussed on more general rapid adaptation methods in neural networks and the direct initial study of how unit level information may be used to facilitate this. Prior to this I completed an MSc in AI an the University of Edinburgh and a BSc in Mathematics at Imperial College London.
You can contact me at any of the social links on the sidebar or email me at ianxmasonsite at gmail dot com.