Functional Regularisation for Continual Learning with Gaussian Processes

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.

17 pages, 7 figures

References in corpus (13)

• Overcoming catastrophic forgetting in neural networks
• Model-Agnostic Meta-Learning for Fast Adaptation of Deep Networks
• Continual Learning Through Synaptic Intelligence
• An Empirical Investigation of Catastrophic Forgetting in Gradient-Based Neural Networks
• Gaussian Processes for Big Data
• Experience Replay for Continual Learning
• Scalable Variational Gaussian Process Classification
• Variational Continual Learning
• Continual Learning with Deep Generative Replay
• Towards Robust Evaluations of Continual Learning
• Continuous Adaptation via Meta-Learning in Nonstationary and Competitive Environments
• Variational Inference for Gaussian Process Modulated Poisson Processes
• Continual Reinforcement Learning with Complex Synapses

Cited by in corpus (43)

• A Review of Uncertainty Quantification in Deep Learning: Techniques, Applications and Challenges
• Understanding the Role of Training Regimes in Continual Learning
• Coresets via Bilevel Optimization for Continual Learning and Streaming
• Domain-Incremental Continual Learning for Mitigating Bias in Facial Expression and Action Unit Recognition
• Continual Deep Learning by Functional Regularisation of Memorable Past
• Continual Learning and Catastrophic Forgetting
• Online Coreset Selection for Rehearsal-based Continual Learning
• Orthogonal Gradient Descent for Continual Learning
• Class-incremental Learning with Pre-allocated Fixed Classifiers
• Federated Continual Learning with Weighted Inter-client Transfer
• Continual Learning in Low-rank Orthogonal Subspaces
• Continual Learning with Adaptive Weights (CLAW)
• Improved Schemes for Episodic Memory-based Lifelong Learning
• Laplace Redux -- Effortless Bayesian Deep Learning
• Strain energy density as a Gaussian process and its utilization in stochastic finite element analysis: application to planar soft tissues
• GP-Tree: A Gaussian Process Classifier for Few-Shot Incremental Learning
• Optimization and Generalization of Regularization-Based Continual Learning: a Loss Approximation Viewpoint
• Linear Mode Connectivity in Multitask and Continual Learning
• Lifelong Policy Gradient Learning of Factored Policies for Faster Training Without Forgetting
• Powerpropagation: A sparsity inducing weight reparameterisation
• Memory-Efficient Semi-Supervised Continual Learning: The World is its Own Replay Buffer
• Evaluating Online Continual Learning with CALM
• Continual Model-Based Reinforcement Learning with Hypernetworks
• Hessian Aware Low-Rank Perturbation for Order-Robust Continual Learning
• Continual Learning via Online Leverage Score Sampling
• Formalizing the Generalization-Forgetting Trade-off in Continual Learning
• Posterior Meta-Replay for Continual Learning
• Bayesian Structure Adaptation for Continual Learning
• Wide Neural Networks Forget Less Catastrophically
• Efficient Data-Parallel Continual Learning with Asynchronous Distributed Rehearsal Buffers
• Detecting and Adapting to Irregular Distribution Shifts in Bayesian Online Learning
• Task-agnostic Continual Learning with Hybrid Probabilistic Models
• Task-Agnostic Online Reinforcement Learning with an Infinite Mixture of Gaussian Processes
• Variational Auto-Regressive Gaussian Processes for Continual Learning
• Graph-Based Continual Learning
• How do Quadratic Regularizers Prevent Catastrophic Forgetting: The Role of Interpolation
• Conditional Deep Gaussian Processes: multi-fidelity kernel learning
• A good body is all you need: avoiding catastrophic interference via agent architecture search
• Embodiment dictates learnability in neural controllers
• Weighted Ensemble Models Are Strong Continual Learners
• Knowledge-Adaptation Priors
• Lifelong Learning Process: Self-Memory Supervising and Dynamically Growing Networks
• FoCL: Feature-Oriented Continual Learning for Generative Models