Modelling continual learning in humans with Hebbian context gating and exponentially decaying task signals
This work addresses continual learning challenges in AI by mimicking human learning mechanisms, offering a biologically inspired solution to reduce interference in neural networks.
The study tackled the problem of catastrophic forgetting in deep neural networks by introducing computational constraints inspired by primate prefrontal cortex gating, enabling sequential learning of two tasks without interference. The model matched human behavioral data, showing performance differences driven by misestimation of category boundaries.
Humans can learn several tasks in succession with minimal mutual interference but perform more poorly when trained on multiple tasks at once. The opposite is true for standard deep neural networks. Here, we propose novel computational constraints for artificial neural networks, inspired by earlier work on gating in the primate prefrontal cortex, that capture the cost of interleaved training and allow the network to learn two tasks in sequence without forgetting. We augment standard stochastic gradient descent with two algorithmic motifs, so-called "sluggish" task units and a Hebbian training step that strengthens connections between task units and hidden units that encode task-relevant information. We found that the "sluggish" units introduce a switch-cost during training, which biases representations under interleaved training towards a joint representation that ignores the contextual cue, while the Hebbian step promotes the formation of a gating scheme from task units to the hidden layer that produces orthogonal representations which are perfectly guarded against interference. Validating the model on previously published human behavioural data revealed that it matches performance of participants who had been trained on blocked or interleaved curricula, and that these performance differences were driven by misestimation of the true category boundary.