Shi Zenglin, Jing Jie, Sun Ying, Lim Joo-Hwee, Zhang Mengmi
IEEE Trans Neural Netw Learn Syst. 2025 Mar 21;PP. doi: 10.1109/TNNLS.2025.3546269.
In artificial intelligence (AI), generalization refers to a model's ability to perform well on out-of-distribution data related to the given task, beyond the data it was trained on. For an AI agent to excel, it must also possess the continual learning capability, whereby an agent incrementally learns to perform a sequence of tasks without forgetting the previously acquired knowledge to solve the old tasks. Intuitively, generalization within a task allows the model to learn underlying features that can readily be applied to novel tasks, facilitating quicker learning and enhanced performance in subsequent tasks within a continual learning framework. Conversely, continual learning methods often include mechanisms to mitigate catastrophic forgetting, ensuring that knowledge from earlier tasks is retained. This preservation of knowledge over tasks plays a role in enhancing generalization for the ongoing task at hand. Despite the intuitive appeal of the interplay of both abilities, existing literature on continual learning and generalization has proceeded separately. In the preliminary effort to promote studies that bridge both fields, we first present empirical evidence showing that each of these fields has a mutually positive effect on the other. Next, building upon this finding, we introduce a simple and effective technique known as shape-texture consistency regularization (STCR), which caters to continual learning. STCR learns both shape and texture representations for each task, consequently enhancing generalization and thereby mitigating forgetting. Remarkably, extensive experiments validate that our STCR, can be seamlessly integrated with existing continual learning methods, including replay-free approaches. Its performance surpasses these continual learning methods in isolation or when combined with established generalization techniques by a large margin. Our data and source code are available at https://github.com/ZhangLab-DeepNeuroCogLab/distillation-style-cnn.
在人工智能（AI）中，泛化是指模型在与给定任务相关的分布外数据上表现良好的能力，这些数据超出了其训练所使用的数据范围。对于一个AI智能体来说，要表现出色，它还必须具备持续学习能力，即智能体能够逐步学习执行一系列任务，而不会忘记先前获取的用于解决旧任务的知识。直观地说，任务内的泛化使模型能够学习潜在特征，这些特征可以很容易地应用于新任务，从而在持续学习框架内促进后续任务的更快学习和性能提升。相反，持续学习方法通常包括减轻灾难性遗忘的机制，以确保保留早期任务的知识。这种跨任务的知识保留有助于提升当前任务的泛化能力。尽管这两种能力相互作用具有直观的吸引力，但关于持续学习和泛化的现有文献是分开进行研究的。在促进跨这两个领域研究的初步努力中，我们首先给出实证证据，表明这两个领域对彼此都有积极影响。接下来，基于这一发现，我们引入一种简单有效的技术，称为形状-纹理一致性正则化（STCR），它适用于持续学习。STCR为每个任务学习形状和纹理表示，从而增强泛化能力并减轻遗忘。值得注意的是，大量实验验证了我们的STCR可以无缝集成到现有的持续学习方法中，包括无回放方法。它的性能在单独使用或与既定的泛化技术结合使用时，都大大超过了这些持续学习方法。我们的数据和源代码可在https://github.com/ZhangLab-DeepNeuroCogLab/distillation-style-cnn获取。
