We are testing different acceleration schemes and algorithmic variants for manifold learning and reinforcement learning. Our codes primarily employ tensorflow and pytorch and we’re hoping to get timing results on the modern GPU systems in JLSE.