DeepAI AI Chat
Log In Sign Up

Learning where to learn: Gradient sparsity in meta and continual learning

by   Johannes von Oswald, et al.
ETH Zurich

Finding neural network weights that generalize well from small datasets is difficult. A promising approach is to learn a weight initialization such that a small number of weight changes results in low generalization error. We show that this form of meta-learning can be improved by letting the learning algorithm decide which weights to change, i.e., by learning where to learn. We find that patterned sparsity emerges from this process, with the pattern of sparsity varying on a problem-by-problem basis. This selective sparsity results in better generalization and less interference in a range of few-shot and continual learning problems. Moreover, we find that sparse learning also emerges in a more expressive model where learning rates are meta-learned. Our results shed light on an ongoing debate on whether meta-learning can discover adaptable features and suggest that learning by sparse gradient descent is a powerful inductive bias for meta-learning systems.


page 1

page 2

page 3

page 4


La-MAML: Look-ahead Meta Learning for Continual Learning

The continual learning problem involves training models with limited cap...

Testing the Genomic Bottleneck Hypothesis in Hebbian Meta-Learning

Recent work has shown promising results using Hebbian meta-learning to s...

Meta Learning Backpropagation And Improving It

Many concepts have been proposed for meta learning with neural networks ...

Meta Networks

Neural networks have been successfully applied in applications with a la...

Learning to learn with backpropagation of Hebbian plasticity

Hebbian plasticity is a powerful principle that allows biological brains...

The least-control principle for learning at equilibrium

Equilibrium systems are a powerful way to express neural computations. A...