Self-Supervised Visual Representation Learning via Residual Momentum | IEEE Journals & Magazine | IEEE Xplore