Abstract:Enabling a neural network to sequentially learn multiple tasks is of great significance for expanding the applicability of neural networks in real-world applications. However, artificial neural networks face the well-known problem of catastrophic forgetting. What is worse, the degradation of previously learned skills becomes more severe as the task sequence increases, known as the long-term catastrophic forgetting. It is due to two facts: first, as the model learns more tasks, the intersection of the low-error parameter subspace satisfying for these tasks becomes smaller or even does not exist; second, when the model learns a new task, the cumulative error keeps increasing as the model tries to protect the parameter configuration of previous tasks from interference. Inspired by the memory consolidation mechanism in mammalian brains with synaptic plasticity, we propose a confrontation mechanism in which Adversarial Neural Pruning and synaptic Consolidation (ANPyC) is used to overcome the long-term catastrophic forgetting issue. The neural pruning acts as long-term depression to prune task-irrelevant parameters, while the novel synaptic consolidation acts as long-term potentiation to strengthen task-relevant parameters. During the training, this confrontation achieves a balance in that only crucial parameters remain, and non-significant parameters are freed to learn subsequent tasks. ANPyC avoids forgetting important information and makes the model efficient to learn a large number of tasks. Specifically, the neural pruning iteratively relaxes the current task's parameter conditions to expand the common parameter subspace of the task; the synaptic consolidation strategy, which consists of a structure-aware parameter-importance measurement and an element-wise parameter updating strategy, decreases the cumulative error when learning new tasks. Our approach encourages the synapse to be sparse and polarized, which enables long-term learning and memory. ANPyC exhibits effectiveness and generalization on both image classification and generation tasks with multiple layer perceptron, convolutional neural networks, and generative adversarial networks, and variational autoencoder. The full source code is available at https://github.com/GeoX-Lab/ANPyC.

Learning to Modulate Random Weights: Neuromodulation-inspired Neural Networks For Efficient Continual Learning

Progressive Learning without Forgetting

Overcoming Long-Term Catastrophic Forgetting Through Adversarial Neural Pruning and Synaptic Consolidation

Exploring Neuromodulation for Dynamic Learning

Diffusion-based neuromodulation can eliminate catastrophic forgetting in simple neural networks

A brain-inspired algorithm that mitigates catastrophic forgetting of artificial and spiking neural networks with low computational cost

Continual learning under domain transfer with sparse synaptic bursting

Bio-inspired, task-free continual learning through activity regularization

Weight Friction: A Simple Method to Overcome Catastrophic Forgetting and Enable Continual Learning

Continual Learning with Dependency Preserving Hypernetworks

Continual Learning with Deep Artificial Neurons

Continual Learning by Modeling Intra-Class Variation

Incorporating neuro-inspired adaptability for continual learning in artificial intelligence

Attention-Based Structural-Plasticity

SynapNet: A Complementary Learning System Inspired Algorithm With Real-Time Application in Multimodal Perception

Achieving a Better Stability-Plasticity Trade-off via Auxiliary Networks in Continual Learning

Neuromodulators generate multiple context-relevant behaviors in a recurrent neural network by shifting activity flows in hyperchannels

Neuromimetic metaplasticity for adaptive continual learning

Defeating Catastrophic Forgetting via Enhanced Orthogonal Weights Modification

Realistic Continual Learning Approach using Pre-trained Models

Elephant Neural Networks: Born to Be a Continual Learner