Online Learning of a Memory for Learning Rates
This addresses the challenge of accelerating learning processes for robotics applications, though it appears incremental as it builds on existing gradient-based optimizers.
The paper tackles the problem of slow learning in robotics by introducing an online meta-learning algorithm that builds a memory model of optimal learning rates from past gradient behaviors, which speeds up MNIST classification and various learning control tasks in both batch and online settings.
The promise of learning to learn for robotics rests on the hope that by extracting some information about the learning process itself we can speed up subsequent similar learning tasks. Here, we introduce a computationally efficient online meta-learning algorithm that builds and optimizes a memory model of the optimal learning rate landscape from previously observed gradient behaviors. While performing task specific optimization, this memory of learning rates predicts how to scale currently observed gradients. After applying the gradient scaling our meta-learner updates its internal memory based on the observed effect its prediction had. Our meta-learner can be combined with any gradient-based optimizer, learns on the fly and can be transferred to new optimization tasks. In our evaluations we show that our meta-learning algorithm speeds up learning of MNIST classification and a variety of learning control tasks, either in batch or online learning settings.