Inference and deference in machine neurology
Arguments about whether or not machine learning is the right answer commonly address the wrong question. By examining closely what a neurologist actually does, I show that no approach other than machine learning could conceivably do it better, not because of the nature of the solution, but because of the nature of the task. Conventional objections, especially on the grounds of inferential opacity, misconceive both the subject and object of neurology, and have needlessly obstructed advances in the field. I illustrate the point with the task of therapeutic inference in the focally injured brain.
Parashkev Nachev is a neurologist whose research programme is primarily devoted to high-dimensional modelling of the brain, in health and disease.
Date: Tuesday 08 May
Location: Roberts Building, G08, Sir David Davies LT (map)
18.00 – 19.00: Seminar
19.00 – 20.00: Drinks reception