A bi-level algorithm adapts scalarization weights online in vector-valued repeated games to obtain sublinear regret bounds and raise convergence to a preferred equilibrium from roughly 50% to 80%.
Title resolution pending
2 Pith papers cite this work. Polarity classification is still indexing.
2
Pith papers citing it
verdicts
UNVERDICTED 2representative citing papers
Nonlinearly preconditioned gradient flows admit global solutions with sublinear or exponential convergence and are dual to mirror descent, solving an infinite-horizon optimal control problem whose value function is the Bregman divergence of the cost.
citing papers explorer
-
Nonlinearly preconditioned gradient flows
Nonlinearly preconditioned gradient flows admit global solutions with sublinear or exponential convergence and are dual to mirror descent, solving an infinite-horizon optimal control problem whose value function is the Bregman divergence of the cost.