A dynamic checkpointing scheme based on reinforcement learning | IEEE Conference Publication | IEEE Xplore