Speakers
Prof.
Emmanouil-Vasileios Vlatakis-Gkaragkounis
(University of Wisconsin-Madison)
Pucheng Xiong
(University of Wisconsin-Madison)
Description
We study uncoupled learning in quantum games, where each player $i$ updates using only their own payoff information and chooses a density matrix $\rho_i$. Classical results show that such uncoupled learning cannot always converge to Nash equilibrium, and we ask which parts of this picture remain true for quantum strategies. We show that two classical impossibility results still persist: smooth uncoupled dynamics and simple stochastic one-recall rules cannot guarantee universal convergence. However, we also identify a positive case: for any finite quantum game with an isolated quantum nash equilibria, a quantum higher-order payoff-based dynamics can still locally stabilize the equilibria.
| Topic | Theoretical foundations of quantum learning |
|---|
Authors
Prof.
Emmanouil-Vasileios Vlatakis-Gkaragkounis
(University of Wisconsin-Madison)
Pucheng Xiong
(University of Wisconsin-Madison)