Does Continual Learning Equally Forget All Parameters?
Distribution shift (e.g., task or domain shift) in continual learning (CL) usually results in catastrophic forgetting of neural networks. Although it can be alleviated by repeatedly replaying buffered data, the every-step replay is time-consuming. In this paper, we study which modules in neural networks are more prone to forgetting by investigating their training dynamics during CL. Our proposed metrics show that only a few modules are more task-specific and sensitively alter between tasks, while others can be shared across tasks as common knowledge. Hence, we attribute forgetting mainly to the former and find that finetuning them only on a small buffer at the end of any CL method can bring non-trivial improvement. Due to the small number of finetuned parameters, such `Forgetting Prioritized Finetuning (FPF)'' is efficient in computation. We further propose a more efficient and simpler method that entirely removes the every-step replay and replaces them by only $k$-times of FPF periodically triggered during CL. Surprisingly, this `$k$-FPF'' performs comparably to FPF and outperforms the SOTA CL methods but significantly reduces their computational overhead and cost. In experiments on several benchmarks of class- and domain-incremental CL, FPF consistently improves existing CL methods by a large margin, and $k$-FPF further excels in efficiency without degrading the accuracy. We also empirically studied the impact of buffer size, epochs per task, and finetuning modules on the cost and accuracy of our methods.
Code (0)
등록된 구현이 없습니다.
Tasks
AllAttributeContinual LearningSimilar Papers 제목 키워드 기반
Rehearsal-Free Modular and Compositional Continual Learning for Language Models
Continual learning aims at incrementally acquiring new knowledge while not forgetting existing knowledge. To overcome catastrophic forgetting, methods are either rehearsal-based, i.e., store data examples from previous t…
Continual LearningTransfer LearningChaotic Continual Learning
Training a deep neural network requires the model to go over training data for several epochs and update network parameters. In continual learning, this process results in catastrophic forgetting which is one of the core…
Continual LearningUncertainty-guided Continual Learning with Bayesian Neural Networks
Continual learning aims to learn new tasks without forgetting previously learned ones. This is especially challenging when one cannot access data from previous tasks and when the model has a fixed capacity. Current regul…
Continual LearningDyTox: Transformers for Continual Learning with DYnamic TOken eXpansion
Deep network architectures struggle to continually learn new tasks without forgetting the previous tasks. A recent trend indicates that dynamic architectures based on an expansion of the parameters can reduce catastrophi…
Class Incremental LearningContinual LearningDecoderIncremental LearningGroup and Exclusive Sparse Regularization-based Continual Learning of CNNs
We present a regularization-based approach for continual learning (CL) of fixed capacity convolutional neural networks (CNN) that does not suffer from the problem of catastrophic forgetting when learning multiple tasks s…
Continual Learning