This will be a two part talk as follows:
Part #1 Whole brain circuit mapping in mouse and marmoset
We cannot understand how brains work without understanding how they are connected, ie the architecture of neural circuitry. This is an open problem that has recently become a focus of activity. The talk will describe a systematic effort to comprehensively map circuits in the brain of the mouse and the marmoset, at a "mesoscopic" scale defined by microinjections of neuronal tracer substances. The acquisition and analysis of the associated large data sets will be described.
Part #2 Phase transitions in machine learning and distributed control
Phase transition phenomena appear in modern engineering as a consequence of having large numbers of variables or degrees of freedom.Two examples will be discussed, drawn from machine learning and distributed control.