Continual learning (CL) trains NN models incrementally from a continuous stream of tasks. To remember previously learned knowledge, prior studies store old samples over a memory hierarchy and replay them when new tasks arrive. Edge devices that adopt CL to preserve data privacy are typically energy-sensitive and thus require high model accuracy while not compromising energy efficiency, i.e., cost-effectiveness. Our work is the first to explore the design space of hierarchical memory replay-based CL to gain insights into achieving cost-effectiveness on edge devices. We present Miro, a novel system runtime that carefully integrates our insights into the CL framework by enabling it to dynamically configure the CL system based on resource states for the best cost-effectiveness. To reach this goal, Miro also performs online profiling on parameters with clear accuracy-energy trade-offs and adapts to optimal values with low overhead. Extensive evaluations show that Miro significantly outperforms baseline systems we build for comparison, consistently achieving higher cost-effectiveness.

翻译：持续学习（CL）从连续的任务流中增量式训练神经网络模型。为记住先前学到的知识，先前研究将旧样本存储于分层内存中，并在新任务到来时进行回放。采用CL以保护数据隐私的边缘设备通常对能量敏感，因此需要在确保高模型精度的同时不牺牲能效，即实现经济高效。我们的研究首次探索基于分层内存回放的CL设计空间，以实现在边缘设备上的经济高效性。本文提出Miro，一种新型系统运行时，通过使其能够基于资源状态动态配置CL系统以获取最佳经济高效性，将我们的见解精心融入CL框架。为实现此目标，Miro还针对具有明显精度-能量权衡的参数执行在线性能剖析，并以较低开销自适应调整至最优值。广泛评估表明，Miro显著优于我们构建的用于比较的基线系统，持续实现更高的经济高效性。