Learning Representations by Stochastic Meta-Gradient Descent in Neural Networks

TL;DR

This paper introduces crossprop, a novel incremental learning algorithm that uses meta-gradient descent to improve feature reuse in neural networks, enhancing adaptability to new tasks.

Contribution

The paper presents crossprop, a new meta-gradient based algorithm that generalizes backprop and promotes feature reuse in neural networks for incremental learning.

Findings

Crossprop effectively reuses features for new tasks.

Crossprop outperforms backprop in continual learning scenarios.

The algorithm introduces a memory parameter for each weight.

Abstract

Representations are fundamental to artificial intelligence. The performance of a learning system depends on the type of representation used for representing the data. Typically, these representations are hand-engineered using domain knowledge. More recently, the trend is to learn these representations through stochastic gradient descent in multi-layer neural networks, which is called backprop. Learning the representations directly from the incoming data stream reduces the human labour involved in designing a learning system. More importantly, this allows in scaling of a learning system for difficult tasks. In this paper, we introduce a new incremental learning algorithm called crossprop, which learns incoming weights of hidden units based on the meta-gradient descent approach, that was previously introduced by Sutton (1992) and Schraudolph (1999) for learning step-sizes. The final update equation introduces an additional memory parameter for each of these weights and generalizes the backprop update equation. From our experiments, we show that crossprop learns and reuses its feature representation while tackling new and unseen tasks whereas backprop relearns a new feature representation.