There is a long tradition in neuroscience (and other sciences) to gain insights into one system by studying other relevant, more-accessible model systems. Much we have learned about the human brain are from studying brains of monkeys, mice, flies, and other model animals. In this talk, I will demonstrate how we can gain understandings of real neural circuits by studying artificial ones.

In particular, I will focus on how machine learning can be used to recapitulate structural principles observed in the olfactory system. I will also discuss how machine learning can help us understand and discover biologically-plausible plasticity rules. Finally, I will discuss our efforts to study neural mechanisms of recurrent neural networks performing many cognitive tasks.