Learning to acquire novel cognitive tasks with evolution, plasticity and meta-meta-learning
This addresses the challenge of developing more intelligent agents capable of flexible learning for AI systems, though it is incremental as it builds on existing meta-learning and evolutionary approaches.
The researchers tackled the problem of enabling neural networks to autonomously learn novel cognitive tasks, which require contextual information processing beyond simple stimulus-response associations, by evolving networks with plastic connections and neuromodulation over a set of simple cognitive tasks. The result was that these evolved networks could automatically modify their connectivity to acquire a new cognitive task not seen during evolution, using only stimuli and rewards.
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 contextual information that must be adequately acquired, stored and processed. While many meta-learning algorithms can design agents that autonomously learn new tasks, cognitive tasks adds another level of learning and memory to typical ``learning-to-learn'' problems. Here we evolve neural networks, endowed with plastic connections and neuromodulation, over a sizable set of simple cognitive tasks adapted from a computational neuroscience framework. 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 emphasize the importance of carefully considering the multiple learning loops involved in the emergence of intelligent behavior.