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.

Catastrophic Forgetting Meets Negative Transfer: Batch Spectral Shrinkage for Safe Transfer Learning.

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

Transferring Pretrained Networks to Small Data Via Category Decorrelation.

Deep Inhomogeneous Regularization for Transfer Learning

Overcoming Catastrophic Forgetting for Fine-Tuning Pre-trained GANs

Transferability Vs. Discriminability: Batch Spectral Penalization for Adversarial Domain Adaptation.

Reduce Catastrophic Forgetting of Dense Retrieval Training with Teleportation Negatives

Learning After Learning: Positive Backward Transfer in Continual Learning

Characterizing and Avoiding Negative Transfer

An Efficient Strategy for Catastrophic Forgetting Reduction in Incremental Learning

An Improved Dual-Channel Network to Eliminate Catastrophic Forgetting

An Empirical Investigation of Catastrophic Forgetting in Gradient-Based Neural Networks

Towards Accurate Knowledge Transfer Via Target-awareness Representation Disentanglement

Restoring Latent Factors Against Negative Transfer Using Partial-Adaptation Nonnegative Matrix Factorization.

Maintaining Discrimination and Fairness in Class Incremental Learning

An Empirical Investigation of Catastrophic Forgeting in Gradient-Based Neural Networks.

Sparsity and Heterogeneous Dropout for Continual Learning in the Null Space of Neural Activations

Flattening Sharpness for Dynamic Gradient Projection Memory Benefits Continual Learning

Alleviating Representational Shift for Continual Fine-tuning.

The unreasonable effectiveness of Batch-Norm statistics in addressing catastrophic forgetting across medical institutions

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