I started my journey with Machine Learning at 2018 when during my MRes course I began to explore machine learning tools and paradigms. Currenty I am in the process of obtaining my Ph.D. at Edinburgh University within CDT in Biomedical AI scheme. Under supervision of Dr Kianoush Nazarpour and Dr Agamemnon Krasoulis I investigate active learning for upper limb prosthetics. The objective of my work includes developing architecture for deep active learning to include human in the loop: by identifying low confidence signals annotated by human oracle in iterative manner to define personalised feature representation. My intention is also to establish an explainable feature space for motor-sensory feedback for users derived from dimensionality reduction approaches. Further on, to apply science to real life scenarios we develop a platform for real-time calibration: feedback interface, motor classification. Using fusion of machine learning paradigms and newly developed techniques during my research, I will enable the framework to evolve over time and adapt to human behaviour (continuous learning) and provide a responsive and interpretable system with co-adaptive and synchronous learning between a machine and a human.
Co-adaptive Human-Machine Learning