Coherent Swap Regret and Channel-Proof Learning
Researchers have introduced a new metric called "coherent swap regret" to evaluate stability in quantum games, which accounts for local quantum operations beyond simple behavior replacement. An algorithm using entropic mirror ascent on the CPTP Choi slice achieves a regret of $O(\sqrt{dT\log d})$. This work establishes a three-level deviation-class landscape, showing that while unitary deviations have zero minimax regret, deterministic measurement-and-preparation channels necessitate significant regret, highlighting the challenge posed by non-unital recommendations. AI