Functional Regularisation for Continual Learning with Gaussian Processes
Preprint
Abstract
We introduce a framework for Continual Learning (CL) based on Bayesian inference over the function space rather than the parameters of a deep neural network. This method, referred to as functional regularisation for Continual Learning, avoids forgetting a previous task by constructing and memorising an approximate posterior belief over the underlying task-specific function. To achieve this we rely on a Gaussian process obtained by treating the weights of the last layer of a neural network as random and Gaussian distributed. Then, the training algorithm sequentially encounters tasks and constructs posterior beliefs over the task-specific functions by using inducing point sparse Gaussian process methods. At each step a new task is first learnt and then a summary is constructed consisting of (i) inducing inputs -- a fixed-size subset of the task inputs selected such that it optimally represents the task -- and (ii) a posterior distribution over the function values at these inputs. This summary then regularises learning of future tasks, through Kullback-Leibler regularisation terms. Our method thus unites approaches focused on (pseudo-)rehearsal with those derived from a sequential Bayesian inference perspective in a principled way, leading to strong results on accepted benchmarks.Keywords
All Related Versions
This publication has 6 references indexed in Scilit:
- Learning without ForgettingIEEE Transactions on Pattern Analysis and Machine Intelligence, 2017
- PowerPlay: Training an Increasingly General Problem Solver by Continually Searching for the Simplest Still Unsolvable ProblemFrontiers in Psychology, 2013
- Kernels for Vector-Valued Functions: A ReviewFoundations and Trends in Machine Learning, 2012
- Sparse On-Line Gaussian ProcessesNeural Computation, 2002
- Catastrophic Forgetting and the Pseudorehearsal Solution in Hopfield-type NetworksConnection Science, 1998
- Catastrophic Forgetting, Rehearsal and PseudorehearsalConnection Science, 1995