Topic: learning times

We present a novel approach using curriculum learning to identify principles by which a neural circuit learns to perform different tasks. Previous work has focused on how curricula can be designed to improve learning of a model on particular tasks. We consider the inverse problem:what can a curriculum tell us about how a learning system acquired a task? Using recurrent neural networks (RNNs) and commonly used tasks in experimental neuroscience, we demonstrate that curricula can be used to differentiate learning principles, with target-based and a representation-based loss functions as use case