Search papers, labs, and topics across Lattice.
1
0
3
0
Controller performance in edge clusters can swing dramatically based on workload intensity, with a deep reinforcement-learning approach losing to a simple heuristic by 29 percentage points under heavy load.
