Learning to acquire novel cognitive tasks with evolution, plasticity and meta-meta-learning

A hallmark of intelligence is the ability to autonomously learn new flexible, cognitive behaviors - that is, behaviors where the appropriate action depends not just on immediate stimuli (as in simple reflexive stimulus-response associations), but on memorized contextual information that must be adequately acquired, stored and processed. Artificial agents can learn such cognitive tasks with external, human-designed meta-learning (``learning-to-learn'') algorithms. By contrast, animals are able to pick up such cognitive tasks automatically, from stimuli and rewards alone, through the operation of their own internal machinery: Evolution has endowed animals with the ability to automatically acquire novel cognitive tasks, including tasks never seen during evolution. Can we harness this process to generate artificial agents with such abilities? Here we evolve neural networks, endowed with plastic connections and neuromodulation, over a sizable set of simple cognitive tasks based on a framework from computational neuroscience. The resulting evolved networks can automatically modify their own connectivity to acquire a novel simple cognitive task, never seen during evolution, from stimuli and rewards alone, through the spontaneous operation of their evolved neural organization and plasticity system. Our results suggest that carefully considering the multiplicity of learning loops involved in intelligent behavior may prove crucial for efficient cognitive learning