The goal of continual learning is to provide intelligent agents that are capable of learning continually a sequence of tasks using the knowledge obtained from previous tasks while performing well on prior tasks. However, a key challenge in this continual learning paradigm is catastrophic forgetting, namely adapting a model to new tasks often leads to severe performance degradation on prior tasks. Current memory-based approaches show their success in alleviating the catastrophic forgetting problem by replaying examples from past tasks when new tasks are learned. However, these methods are infeasible to transfer the structural knowledge from previous tasks i.e., similarities or dissimilarities between different instances. Furthermore, the learning bias between the current and prior tasks is also an urgent problem that should be solved. In this work, we propose a new method, named Online Continual Learning via the Knowledge Invariant and Spread-out Properties (OCLKISP), in which we constrain the evolution of the embedding features via Knowledge Invariant and Spread-out Properties (KISP). Thus, we can further transfer the inter-instance structural knowledge of previous tasks while alleviating the forgetting due to the learning bias. We empirically evaluate our proposed method on four popular benchmarks for continual learning: Split CIFAR 100, Split SVHN, Split CUB200 and Split Tiny-Image-Net. The experimental results show the efficacy of our proposed method compared to the state-of-the-art continual learning algorithms.
翻译:持续学习的目的是提供智能动力,能够利用从以往工作中获得的知识不断学习一系列任务,同时很好地完成以往任务。然而,这种持续学习模式中的一个关键挑战是灾难性的忘记,即将模式适应新任务,往往导致以往任务的业绩严重退化。当前基于记忆的方法表明,它们成功地减轻了灾难性的遗忘问题,在学习新任务时,它们重复了过去任务的例子。然而,这些方法无法从以往任务中传递结构性知识,即不同实例之间的相似或差异。此外,当前和以往任务之间的学习偏差也是一个紧迫的问题,应当加以解决。在这项工作中,我们提出了一个新方法,名为“在线持续学习”,通过知识不易和扩展属性(OCLKISP),通过知识不易和扩展属性(KISP)来限制嵌入特性的演进。因此,我们可以进一步转移以往任务之间的内部结构知识知识,同时减轻我们的学习偏差。我们从经验上评估了我们提议的关于四个通用的C-N持续学习基准的方法,将Sliveral-Slivering Streal-S-I Sliveral-Sligal-Slistal-Sligal-Sligal-Sligal-Slipal sh-Slipal-Slistal-Slipal-Slist-Sl-Sl-Slipal-Slipal-Slipleglegal-S-S-S-S-Sl-S-S-S-S-Slipal-S-Slipal-Sl-S-S-S-S-S-Sl-Sl-Sl-S-S-Sl-S-S-S-Sl-S-S-S-S-S-S-S-S-Sl-S-S-S-S-S-S-S-S-S-S-S-S-S-S-S-S-S-S-S-S-I-Sl-Sl-I-I-S-S-S-S-S-S-S-S-S-S-S-S-S-S-S-S-S-S-S-S-S-S-S-S-S-S-S-S-S-S-S