Online Knowledge Acquisition with Selective Inherited Model
Xiacong Du · OSTI OAI (U.S. Department of Energy Office of Scientific and Technical Information) · 2020
Continual learning, which updates the machine learning model from streamed data, is increasingly needed in the dynamic systems.Such a scenario requires both the preservation of previous knowledge, as well as the adaptation to new observations, with high computing efficiency and memory usage at the edge.Previous approaches attempt to learn the knowledge class by class from scratch, using either regularization based or memory replay-based methods.However, they still suffer from severe accuracy drop, a.k.a catastrophic forgetting, during this incremental process; moreover, as the entire model is involved in each updating, their computation cost is too expensive for edge computing.In this work, we propose a novel braininspired paradigm named acquisitive learning (AL).Different from previous approaches that focus only on model adaptation, AL emphasizes the importance of both knowledge inheritance and acquisition: the knowledge is first pre-trained and selected in the cloud (the inherited model and selection) and then adapted to new knowledge (the acquisition).The quality of the inherited model is monitored by the landscape of the loss function, while the acquisition is realized by segmented training.The combination of both steps reduces accuracy drop by >10X on the CIFAR-100 dataset.Furthermore, AL benefits edge computing with 5X reduction in latency per training image on FPGA prototype and 150X reduction in training FLOPs.