arXiv:2609. 16751v1 Announce Type: cross Abstract: We give deterministic and uncoupled learning dynamics for finite multiplayer general-sum games under full-information feedback that achieve constant individual swap regret, independent of the horizon $T$.
We give deterministic and uncoupled learning dynamics for finite multiplayer general-sum games under full-information feedback that achieve constant individual swap regret, independent of the horizon...
arXiv:2609.16751v2 Announce Type: replace-cross Abstract: We give deterministic and uncoupled learning dynamics for finite multiplayer general-sum games under full-information feedback that achieve c...
arXiv:2608. 04149v1 Announce Type: cross Abstract: Swap regret governs the rate at which uncoupled learning dynamics converge to correlated equilibria in multiplayer general-sum games.
arXiv:2609.22839v1 Announce Type: cross Abstract: Can simple learning rules keep their regret bounded in self-play? Recent work achieves constant regret bounds through modified regularization and hig...
The paper presents an uncoupled learning algorithm, higher-order optimism with discounting (HOOD), for arbitrary N-player normal form games with up to K actions per player. HOOD achieves an individual regret bound of O(N³ log² K) uniformly over the play horizon by combining a discounted (N+1)-th order predictor with entropic regularization over a lifted strategy space. This design mitigates large oscillations in play, addressing a key challenge in prior attempts to attain constant regret in general games.
arXiv:2607. 23333v1 Announce Type: cross Abstract: We revisit the regret loss framework introduced in Park et al.
arXiv:2609.21976v1 Announce Type: cross Abstract: We introduce Multiplicatively Optimistic Regret Matching (MORM), an uncoupled learning rule for finite general-sum games. Under simultaneous full-inf...
arXiv:2608. 31166v1 Announce Type: new Abstract: Uncoupled no-regret dynamics provide a decentralized route to equilibrium, but prior guarantees for individual regret retain a polylogarithmic dependence on the horizon.
arXiv:2606. 29533v1 Announce Type: cross Abstract: We study the problem of forecasting for an arbitrary number of downstream agents with unknown objectives, each of whom best responds to the forecaster's predictions.
The paper introduces Decision‑Relevant Fresh Comparison (DRFC), a method for decentralized bandit systems with heterogeneous agents whose local reward changes may not affect the global best action. DRFC gathers balanced samples from all agents and only switches the common best arm when fresh global evidence indicates a change, yielding a dynamic regret bound that does not depend on the number of local changes. An anytime‑valid sliding‑window extension further handles gradual drift, and experiments on synthetic, semi‑real, and MovieLens‑1M data demonstrate that DRFC ignores decision‑irrelevant local changes while the extension avoids false switches.
The paper introduces a straightforward framework that transforms dynamic regret minimization into switching regret minimization by constructing an unbiased random sequence for any comparator sequence. Using this reduction, the authors derive dynamic regret bounds for strongly convex and exp-concave losses of “~O(T^{1/3}P_T^{2/3})” and for general convex losses of “O(√{T(1+P_T)})”, matching known minimax optimal results. The approach leverages off-the-shelf switching regret algorithms and controlled variance to achieve these bounds.
arXiv:2608. 09389v1 Announce Type: cross Abstract: This note aims to serve as an entry point to the literature on learning in games, a topic with significant theoretical appeal and a wide range of applications -- from machine learning and data science to economics and beyond.