Perceptual learning, roving and the unsupervised bias.

33 citations recorded by the source corpus.

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

Synaptic weight decay with selective consolidation enables fast learning without catastrophic forgetting (2019, methodology): “A previous model was explaining the roving phenomenon by an imprecise critic in the context of reinforcement learning (Herzog et al., 2012).”
Comparing continual task learning in minds and machines (2018, background): “For example, it is possible that switching between two reward functions in the interleaved learning pro-vokes a generic processing cost in humans that is lower or absent in the blocked condition (37).”
Deleterious effects of roving on learned tasks. (2014, unknown): “A recent mathematical analysis showed that roving hinders learning for tasks with differing difficulty levels because the critic of reinforcement learning models cannot assign ‘‘reward’’ individually to two similar stimulus types (because of the unsupervised bias; Frémaux, Sprekeler, & Gerstner, 201”
Neuromodulated Spike-Timing-Dependent Plasticity, and Theory of Three-Factor Learning Rules (2016, methodology): “Recently, Potjans et al. (2009, 2011) have used an actor-critic network of leaky integrate-and-fire neurons to solve a simple 5 × 5 gridworld task with spiking neurons.”