Continual Learning of Context-dependent Processing in Neural Networks

Abstract

Deep neural networks (DNNs) are powerful tools in learning sophisticated but fixed mapping rules between inputs and outputs, thereby limiting their application in more complex and dynamic situations in which the mapping rules are not kept the same but changing according to different contexts. To lift such limits, we developed a novel approach involving a learning algorithm, called orthogonal weights modification (OWM), with the addition of a context-dependent processing (CDP) module. We demonstrated that with OWM to overcome the problem of catastrophic forgetting, and the CDP module to learn how to reuse a feature representation and a classifier for different contexts, a single network can acquire numerous context-dependent mapping rules in an online and continual manner, with as few as $\sim$10 samples to learn each. This should enable highly compact systems to gradually learn myriad regularities of the real world and eventually behave appropriately within it.

References in corpus (3)

• Overcoming catastrophic forgetting in neural networks
• Progress & Compress: A scalable framework for continual learning
• Alleviating catastrophic forgetting using context-dependent gating and synaptic stabilization

Cited by in corpus (15)

• Class-Incremental Learning: A Survey
• Lifelong Incremental Reinforcement Learning with Online Bayesian Inference
• Continual Learning and Catastrophic Forgetting
• One-shot Federated Learning without Server-side Training
• Multi-View Class Incremental Learning
• Beneficial Perturbation Network for designing general adaptive artificial intelligence systems
• Class Gradient Projection For Continual Learning
• Recursive Least-Squares Estimator-Aided Online Learning for Visual Tracking
• Quantum continual learning on a programmable superconducting processor
• Continual Learning with Neuromorphic Computing: Foundations, Methods, and Emerging Applications
• Parameter-Efficient Continual Fine-Tuning: A Survey
• CL-BioGAN: Biologically-Inspired Cross-Domain Continual Learning for Hyperspectral Anomaly Detection
• A Brain-Inspired Sequence Learning Model based on a Logic
• Unsupervised Replay Strategies for Continual Learning with Limited Data
• SEDEG:Sequential Enhancement of Decoder and Encoder's Generality for Class Incremental Learning with Small Memory