SIESTA：带睡眠阶段的高效在线持续学习
SIESTA: Efficient Online Continual Learning with Sleep
AI总结：
本文提出基于 wake/sleep 框架的 SIESTA 持续学习方法，通过无重放、无反向传播的在线更新与睡眠阶段受限重放巩固，在单 GPU 上 2 小时内完成 ImageNet-1K 学习并达到离线性能。
摘要：
在有监督持续学习中，深度神经网络（DNN）需要根据不断增长的数据流进行更新。与数据经过打乱的离线设置不同，我们不能对数据流作出任何分布假设。理想情况下，为了计算效率，只需对数据集进行一次遍历。然而，现有方法并不完善，作出了许多在真实世界应用中无法成立的假设，同时也未能提升计算效率。本文提出一种名为 SIESTA 的新型持续学习方法，它基于用于训练的 wake/sleep（清醒/睡眠）框架，与设备端学习的需求高度契合。SIESTA 的主要目标是推进计算高效的持续学习，使 DNN 能够以少得多的时间和能量得到高效更新。SIESTA 的核心创新包括：1）在 wake 阶段使用无重放、无反向传播且由数据驱动的网络更新规则实现快速在线更新；2）在 sleep 阶段使用计算受限的重放策略来加速记忆巩固。为了提升内存效率，SIESTA 借鉴 REMIND 的内存索引，采用了潜在重放。与 REMIND 及先前方法相比，SIESTA 的计算效率高得多，可在单块 GPU 上于 2 小时内完成 ImageNet-1K 上的持续学习；此外，在无数据增强设置下，其性能与离线学习器相当，这是推动持续学习在真实世界应用中落地的关键里程碑。
英文摘要：
In supervised continual learning, a deep neural network (DNN) is updated with an ever-growing data stream. Unlike the offline setting where data is shuffled, we cannot make any distributional assumptions about the data stream. Ideally, only one pass through the dataset is needed for computational efficiency. However, existing methods are inadequate and make many assumptions that cannot be made for real-world applications, while simultaneously failing to improve computational efficiency. In this paper, we propose a novel continual learning method, SIESTA based on wake/sleep framework for training, which is well aligned to the needs of on-device learning. The major goal of SIESTA is to advance compute efficient continual learning so that DNNs can be updated efficiently using far less time and energy. The principal innovations of SIESTA are: 1) rapid online updates using a rehearsal-free, backpropagation-free, and data-driven network update rule during its wake phase, and 2) expedited memory consolidation using a compute-restricted rehearsal policy during its sleep phase. For memory efficiency, SIESTA adapts latent rehearsal using memory indexing from REMIND. Compared to REMIND and prior arts, SIESTA is far more computationally efficient, enabling continual learning on ImageNet-1K in under 2 hours on a single GPU; moreover, in the augmentation-free setting it matches the performance of the offline learner, a milestone critical to driving adoption of continual learning in real-world applications.
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。
