AdaCL：自适应持续学习
AdaCL:Adaptive Continual Learning
AI总结：
针对类增量学习中超参数固定、难以适应任务新颖性与难度变化的问题，提出基于贝叶斯优化的 AdaCL，为每个新任务自动调节学习率、正则化强度和内存大小，从而改善准确率、遗忘与内存表现。
摘要：
类增量学习旨在更新深度分类器以学习新类别，同时保持或提高其在此前观察到类别上的准确率。防止遗忘已学习类别的常见方法包括对神经网络更新进行正则化以及在内存中存储样本，这些方法涉及学习率、正则化强度或样本数量等超参数。然而，这些超参数通常只在开始时进行调优，随后在整个学习过程中保持固定，忽视了新遇到任务可能具有不同程度新颖性或难度的事实。本研究考察了类增量学习中超参数“自适应性”的必要性：即根据当前新任务的属性动态调整学习率、正则化强度和内存大小等超参数的能力。我们提出 AdaCL，这是一种基于贝叶斯优化的方法，能够针对每个学习任务自动且高效地确定这些参数的最优值。我们表明，在每个新任务上调整超参数能够带来准确率、遗忘程度和内存方面的改善。代码可在 https://github.com/ElifCerenGokYildirim/AdaCL 获取。
英文摘要：
Class-Incremental Learning aims to update a deep classifier to learn new categories while maintaining or improving its accuracy on previously observed classes. Common methods to prevent forgetting previously learned classes include regularizing the neural network updates and storing exemplars in memory, which come with hyperparameters such as the learning rate, regularization strength, or the number of exemplars. However, these hyperparameters are usually only tuned at the start and then kept fixed throughout the learning sessions, ignoring the fact that newly encountered tasks may have varying levels of novelty or difficulty. This study investigates the necessity of hyperparameter `adaptivity' in Class-Incremental Learning: the ability to dynamically adjust hyperparameters such as the learning rate, regularization strength, and memory size according to the properties of the new task at hand. We propose AdaCL, a Bayesian Optimization-based approach to automatically and efficiently determine the optimal values for those parameters with each learning task. We show that adapting hyperpararmeters on each new task leads to improvement in accuracy, forgetting and memory. Code is available at https://github.com/ElifCerenGokYildirim/AdaCL.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
