arXiv research

New method estimates state-action stationary distribution for better off-policy policy evaluation.

Study stationary measures and orbit closures for non-abelian actions on surfaces.

In order to apply variational methods to the action functional for geodesics of a stationary spacetime, some hypotheses, useful to obtain classical Palais-Smale condition, are commonly used: pseudo-coercivity, bounds on certain coefficients of the metric, etc. We prove that these technical assumptions admit a natural i…

New algorithm tracks changes in infinite action space rewards.

KeRNS tackles non-stationary reinforcement learning in metric spaces.

NVMDP framework tackles non-stationary MDPs with varying discount rates.

We prove an equivariant deformation result for Hamiltonian stationary Lagrangian submanifolds of a Kahler manifold, with respect to deformations of its metric and almost complex structure that are compatible with an isometric Hamiltonian group action. This yields existence of Hamiltonian stationary Lagrangian submanifo…

New method uses SMC for Bayesian MABs with time-varying, nonlinear rewards.

MetaCURL tackles non-stationary MDPs with optimal dynamic regret.

Non-stationary reinforcement learning is challenging due to the complexity of updating value functions.

Optimal algorithms for non-stationary bandits minimize regret over time.

Formula calculates Riemann-Roch number for singular symplectic quotients.

Observer learns optimal policy from learner's actions without rewards.

We obtain an explicit parametrization of stationary discs glued to some Levi non-degenerate hypersurfaces. These discs form a family which is invariant under the action of biholomorphisms. We use this parametrization to construct a local circular representation of these hypersurfaces. As a corollary, we get the uniquen…

Predictive sampling improves on Thompson sampling for non-stationary bandit environments.

The paper studies limit sets on $\mathbb{P}(\mathbb{R}^3)$ using stationary measures.

Study measures rigidity for random walks and flows via generalized u-Gibbs states.

We prove some ergodic-theoretic rigidity properties of the action of SL(2,R) on moduli space. In particular, we show that any ergodic measure invariant under the action of the upper triangular subgroup of SL(2,R) is supported on an invariant affine submanifold. The main theorems are inspired by the results of several a…

We show that most homogeneous Anosov actions of higher rank Abelian groups are locally smoothly rigid (up to an automorphism). This result is the main part in the proof of local smooth rigidity for two very different types of algebraic actions of irreducible lattices in higher rank semisimple Lie groups: (i) the Anosov…

New algorithm tackles non-stationary reinforcement learning with general function approximation.

New method stabilizes FQE by reweighting Bellman targets.

The paper suggests asset prices follow physical laws, allowing for accurate price movement forecasts.

Study tackles non-stationary bandit convex optimization with new algorithms.

New method improves stability of soft FQI for offline RL.

Internal Lagrangians derived from variational principles.

Memory-based models can learn to approximate Bayes-optimal predictors for non-stationary data.

Simple anomaly detector adapts to non-stationary data.

New algorithm optimizes multi-objective outcomes in uncertain environments.

Study tackles OPE in confounded settings, estimating policy value from proxies.

New RL algorithm tackles non-stationary environments with flexible policy updates.

The existence of stationary Markov perfect equilibria in stochastic games is shown under a general condition called "(decomposable) coarser transition kernels". This result covers various earlier existence results on correlated equilibria, noisy stochastic games, stochastic games with finite actions and state-independe…

We consider the reduced Allen-Cahn action functional, which appears as the sharp interface limit of the Allen-Cahn action functional and can be understood as a formal action functional for a stochastically perturbed mean curvature flow. For suitable evolutions of generalized hypersurfaces this functional consists of th…

Develops a new method for optimizing policies in hierarchical models.

Reward suffices for convex MDPs, expanding RL to new problems.

A new algorithm for non-stationary linear bandits with improved regret bound.

A new method optimizes in nonstationary environments with many arms efficiently.

Study extreme values of stable random fields on geometric spaces.

The paper studies critical points and flows of a $G_2$-Hilbert functional on manifolds with circle actions.

Paper introduces Decentralized Non-stationary Competing Bandits ( exttt{DNCB}) for dynamic matching markets.

Two randomized algorithms improve performance in non-stationary linear bandits.

A new kernel improves Gaussian process performance for non-stationary data.

New findings on universal learning in contextual bandits with adversarial rewards.

Formula derived for spectral determinant of sphere with conical singularities.

Unified approach for non-stationary linear bandits with dynamic regret.

The paper deals with the Weyl equation which is the massless Dirac equation. We study the Weyl equation in the stationary setting, i.e. when the spinor field oscillates harmonically in time. We suggest a new geometric interpretation of the stationary Weyl equation, one which does not require the use of spinors, Pauli m…

The second order differential equation $\frac{D\dotγ}{dt}(t) = F_{γ(t)}(\dotγ(t)) - \nabla V(γ(t))$ on a Lorentzian manifold describes, in particular, the dynamics of particles under the action of a electromagnetic field $F$ and a conservative force $-\nabla V$. We provide a first study on the extendability of its solu…

If a differential equation in a Banach manifold is invariant or quasi-invariant under the action of one or more Lie groups, then its stationary points cannot be isolated, so that classical linearized stability theorem does not apply to it. The first main purpose of this paper is to establish a linearized stability theo…

New TS algorithms improve performance in non-stationary multi-armed bandit problems.