arXiv research

PHE adds pseudo-rewards to history to minimize regret in stochastic bandits.

A new bandit algorithm Giro learns to explore by sampling its history with pseudo rewards.

New algorithm improves materials discovery using max K-Armed Bandit.

Logarithmic regret strategies for safe multi-armed bandits with safety risk constraints.

We consider an agent's uncertainty about its environment and the problem of generalizing this uncertainty across observations. Specifically, we focus on the problem of exploration in non-tabular reinforcement learning. Drawing inspiration from the intrinsic motivation literature, we use density models to measure uncert…

A fair policy for hiring candidates from different groups is proposed in a linear contextual bandit problem.

A single pre-trained agent guides feature selection using knockoffs.

Paper shows RL's policy gradient methods can be viewed as supervised learning problems.

New algorithm for RL with horizon-free reward-free exploration for linear MDPs.

Study shows fast rates for inverse reinforcement learning with linear rewards.

Successor Options discovers reusable skills using landmark states.

New algorithm minimizes regret in stochastic linear bandits with perturbed history.

We study, to the best of our knowledge, the first Bayesian algorithm for unimodal Multi-Armed Bandit (MAB) problems with graph structure. In this setting, each arm corresponds to a node of a graph and each edge provides a relationship, unknown to the learner, between two nodes in terms of expected reward. Furthermore, …

Improves neural networks' ability to learn new tasks without forgetting earlier ones.

Characterizes UCB algorithm performance in bandit problems.

A novel bandit problem with delayed arms, showing optimal strategies and lower bounds.

New attack manipulates UCB algorithm, new defense algorithm reduces pseudo-regret.

A method for continual learning using world models in reinforcement learning.

New algorithm robust to probabilistic unbounded adversarial attacks in bandit problems.

SUPE combines unlabeled data with RL to efficiently explore tasks.

Paper proposes a deep RL method for hedging variable annuities, outperforming misspecified models.

We classify pseudo-Riemannian submersions with connected totally geodesic fibres from a real pseudo-hyperbolic space onto a pseudo-Riemannian manifold. Also, we obtain the classification of the pseudo-Riemannian submersions with (para-)complex connected totally geodesic fibres from a (para-)complex pseudo-hyperbolic sp…

The CR analogue of B.-Y. Chen's conjecture on pseudo biharmonic maps will be shown. Pseudo biharmonic, but not pseudo harmonic, isometric immersions with parallel pseudo mean curvature vector fields, will be characterized. Several examples of pseudo biharmonic maps will be given.

Study of pseudo knots, links, and knotoids with braiding and L-moves.

The paper classifies conformal solitons in pseudo-Euclidean spaces.

In this work, we study the pseudo-Riemannian submanifolds of a pseudo-sphere with 1-type pseudo-spherical Gauss map. First, we classify the Lorentzian surfaces in a 4-dimensional pseudo-sphere $\mathbb{S}^4_s(1)$ with index s, $s=1, 2$, and having harmonic pseudo-spherical Gauss map. Then we give a characterization the…

The paper extends knot theory to annular and toroidal pseudo knots.

Extends tangle theory to include undetermined crossings in periodic structures.

The paper studies special surfaces in pseudo-Euclidean space.

Paper proposes DG-ETC for online submodular maximization with stochastic bandit feedback.

In this paper, we give an estimate of sub-Laplacian of Riemannian distance functions in pseudo-Hermitian geometry which plays a similar role as Laplacian comparison theorem in Riemannian geometry, and deduce a prior horizontal gradient estimate of pseudo-harmonic maps from pseudo-Hermitian manifolds to regular balls of…

Meta Pseudo Labels boosts image classification accuracy to 90.2%.

The paper studies prolongations of Lie algebras associated with pseudo $H$-type Lie algebras.

In this paper, we derived biharmonic equations for pseudo-Riemannian submanifolds of pseudo-Riemannian manifolds which includes the biharmonic equations for submanifolds of Riemannian manifolds as a special case. As applications, we proved that a pseudo-umbilical biharmonic pseudo-Riemannian submanifold of a pseudo-Rie…

The paper studies pseudo links in genus g handlebodies, generalizing knot theory.

Paper constructs a HOMFLYPT-type invariant for pseudo links.

New Einstein metrics found on specific Lie algebras.

Paper studies pseudo-projective tensors on warped products.

Study pseudo-Riemannian metrics on Jordan superalgebras.

The paper explores knotoids, pseudo knotoids, braidoids, and pseudo braidoids on the torus.

The orthogonal decomposition of the Webster curvature provides us a way to characterize some canonical metrics on a pseudo-Hermitian manifold. We derive some subelliptic differential inequalities from the Weitzenböck formulas for the traceless pseudo-Hermitian Ricci tensor and the Chern-Moser tensor of Sasakian manifol…

It is well known that the curvature tensor of a pseudo-Riemannian manifold can be decomposed with respect to the pseudo-orthogonal group into the sum of the Weyl conformal curvature tensor, the traceless part of the Ricci tensor and of the scalar curvature. A similar decomposition with respect to the pseudo-unitary gro…

The paper characterizes metallic pseudo-Riemannian manifolds using conjugate connections and tensor structures.

The paper proves Liouville theorems for holomorphic maps on pseudo-Hermitian manifolds.

Reward hacking exploits misspecified rewards, affecting agent capabilities and true performance.

The paper aims to initiate a systematic study of conformal mappings between Finsler spacetimes and, more generally, between pseudo-Finsler spaces. This is done by extending several results in pseudo-Riemannian geometry which are necessary for field-theoretical applications and by proposing a technique which reduces a s…

Study on pseudo-Riemannian algebraic Ricci solitons in 4D Lie groups.

Paper introduces PRMs to learn non-Markovian stochastic rewards for reinforcement learning.