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.

Reducing Catastrophic Forgetting With Associative Learning: A Lesson From Fruit Flies

Algorithmic insights on continual learning from fruit flies

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

Incorporating neuro-inspired adaptability for continual learning in artificial intelligence

Association: Remind Your GAN not to Forget

Memory Aware Synapses: Learning what (not) to forget

Learn to Grow: A Continual Structure Learning Framework for Overcoming Catastrophic Forgetting

Understanding Catastrophic Forgetting and Remembering in Continual Learning with Optimal Relevance Mapping

Fortuitous Forgetting in Connectionist Networks

Attention-Based Structural-Plasticity

Contributions by metaplasticity to solving the Catastrophic Forgetting Problem

Behavioral Experiments for Understanding Catastrophic Forgetting

Reducing Catastrophic Forgetting in Self Organizing Maps with Internally-Induced Generative Replay

A Behavioral Paradigm to Study the Persistence of Reward Memory Extinction in Drosophila.

Reducing catastrophic forgetting when evolving neural networks

A Methodology-Oriented Study of Catastrophic Forgetting in Incremental Deep Neural Networks

Measuring Catastrophic Forgetting in Neural Networks

Neuronal circuit mechanisms of competitive interaction between action-based and coincidence learning

AFEC: Active Forgetting of Negative Transfer in Continual Learning

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

Switch-like and persistent memory formation in individual Drosophila larvae