Description: Towards general-purpose machine learning
Our focus is computational descriptions that are implementable .
We created an architecture with hemispheric specialization to better understand the brain and uncover new principles for AI
We found that replay improves continual few-shot learning significantly, for learning classes as well as specific instances