Mitigating Forgetting in Online Continual Learning with Neuron Calibration
Inspired by human intelligence, the research on online continual learning aims to push the limits of the machine learning models to constantly learn from sequentially encountered tasks, with the data from each task being observed in an online fashion. Though recent studies have achieved remarkable progress in improving the online continual learning performance empowered by the deep neural networks-based models, many of today's approaches still suffer a lot from catastrophic forgetting, a persistent challenge for continual learning. In this paper, we present a novel method which attempts to mitigate catastrophic forgetting in online continual learning from a new perspective, i.e., neuron calibration. In particular, we model the neurons in the deep neural networks-based models as calibrated units under a general formulation. Then we formalize a learning framework to effectively train the calibrated model, where neuron calibration could give ubiquitous benefit to balance the stability and plasticity of online continual learning algorithms through influencing both their forward inference path and backward optimization path. Our proposed formulation for neuron calibration is lightweight and applicable to general feed-forward neural networks-based models. We perform extensive experiments to evaluate our method on four benchmark continual learning datasets. The results show that neuron calibration plays a vital role in improving online continual learning performance and our method could substantially improve the state-of-the-art performance on all~the~evaluated~datasets.
Code (0)
등록된 구현이 없습니다.
Tasks
Continual LearningSimilar Papers 제목 키워드 기반
Mitigating Forgetting in Online Continual Learning with Neuron Calibration
Inspired by human intelligence, the research on online continual learning aims to push the limits of the machine learning models to constantly learn from sequentially encountered tasks, with the data from each task being…
Continual LearningSAMix: Calibrated and Accurate Continual Learning via Sphere-Adaptive Mixup and Neural Collapse
While most continual learning methods focus on mitigating forgetting and improving accuracy, they often overlook the critical aspect of network calibration, despite its importance. Neural collapse, a phenomenon where las…
Continual LearningContinual Learning for Seq2Seq Generations with Transformer Calibration
Conventional NLP generation models are trained offline with a given dataset for a particular task, which is referred to as isolated learning. Research on sequence-to-sequence language generation aims to study continual l…
Continual LearningLifelong learningParaphrase GenerationText GenerationConsistency is the key to further mitigating catastrophic forgetting in continual learning
Deep neural networks struggle to continually learn multiple sequential tasks due to catastrophic forgetting of previously learned tasks. Rehearsal-based methods which explicitly store previous task samples in the buffer …
Continual LearningSelf-Supervised LearningLearning Bayesian Sparse Networks with Full Experience Replay for Continual Learning
Continual Learning (CL) methods aim to enable machine learning models to learn new tasks without catastrophic forgetting of those that have been previously mastered. Existing CL approaches often keep a buffer of previous…
Continual LearningKnowledge DistillationOpen Knowledge Graph Canonicalization