arXiv research

AE-DQN learns to eliminate sub-optimal actions in complex RL environments.

A scheme robust to action erasures improves MAB performance.

Robust algorithm optimizes corrupted Gaussian process bandits.

We study an original problem of pure exploration in a strategic bandit model motivated by Monte Carlo Tree Search. It consists in identifying the best action in a game, when the player may sample random outcomes of sequentially chosen pairs of actions. We propose two strategies for the fixed-confidence setting: Maximin…

Paper tackles online learning in large MDPs with low Bellman rank using AVE algorithm.

We initiate the study of multi-stage episodic reinforcement learning under adversarial corruptions in both the rewards and the transition probabilities of the underlying system extending recent results for the special case of stochastic bandits. We provide a framework which modifies the aggressive exploration enjoyed b…

New algorithm reduces regret in stochastic linear bandits with heteroscedastic noise.

New algorithm identifies best intervention without graph knowledge.

Partial connections are (singular) differential systems generalizing classical connections on principal bundles, yielding analogous decompositions for manifolds with nonfree group actions. Connection forms are interpreted as maps determining projections of the tangent bundle onto the partial connection; this approach e…

The paper tackles fair sequential decision making with biased linear bandit feedback.

Paper tackles action selection in deep RL, proposing a data-driven approach.

Study lenient regret and good-action identification in Gaussian process bandits.

A federated learning algorithm tackles unknown contexts in multi-arm bandits.

Neural eliminators reduce unreliable classification by eliminating improbable classes.

Proposes a flexible tournament design combining knockout and round-robin.

We establish a new connection between value and policy based reinforcement learning (RL) based on a relationship between softmax temporal value consistency and policy optimality under entropy regularization. Specifically, we show that softmax consistent action values correspond to optimal entropy regularized policy pro…

Improved elimination strategies for adaptive bandit identification reduce sample complexity and computational burden.

We generalise to the $\mathbb{Z}_2$-graded set-up a practical method for inspecting the (non)removability of parameters in zero-curvature representations for partial differential equations (PDEs) under the action of smooth families of gauge transformations. We illustrate the generation and elimination of parameters in …

New algorithm eliminates arms to minimize regret in complex bandit problems.

Tensor variable elimination for plated factor graphs enables exact inference in models with repeated structure.

A new ML method predicts long-time-step molecular dynamics, preserving symplectic and time-reversible properties.

A new method for RL with continuous actions improves stability and scalability.

CoverNet predicts urban driving trajectories using diverse sets of possible actions.

We introduce Neural Choice by Elimination, a new framework that integrates deep neural networks into probabilistic sequential choice models for learning to rank. Given a set of items to chose from, the elimination strategy starts with the whole item set and iteratively eliminates the least worthy item in the remaining …

Aims to eliminate domain bias in authentication without domain labels.

New BE dimension measure reveals rich RL problems with sample-efficient algorithms.

Approximate models help RL by reducing policy search space.

We develop a model to study the role of rationality in economics and biology. The model's agents differ continuously in their ability to make rational choices. The agents' objective is to ensure their individual survival over time or, equivalently, to maximize profits. In equilibrium, however, rational agents who maxim…

This paper proves adding neurons eliminates all bad local minima in deep learning.

We simplify Khovanov homology for torus braids using Gaussian elimination.

AMBER method selects features efficiently using autoencoders and model-based elimination.

Logic approach finds real singularities in differential equations.

We develop an approach for feature elimination in statistical learning with kernel machines, based on recursive elimination of features.We present theoretical properties of this method and show that it is uniformly consistent in finding the correct feature space under certain generalized assumptions.We present four cas…

Method models other agents' behaviors without requiring direct observation.

Optimization geometrodynamics simplifies adaptive optimizer dynamics.

We extend several techniques and theorems from geometric group theory so that they apply to geometric actions on arbitrary proper metric ARs (absolute retracts). A second way that we generalize earlier results is by eliminating freeness requirements often placed on the group actions. In doing so, we allow for groups wi…

AlphaForgeBench evaluates LLMs as quantitative researchers, not trading agents, to address instability in financial decision-making.

A new algorithm improves recommendation systems by considering repeated exposure to actions.

Two new feature selection algorithms improve on RFE.

Let D be an irreducible lattice in a connected, semisimple Lie group G with finite center. Assume that the real rank of G is at least two, that G/D is not compact, and that G has more than one noncompact simple factor. We show that D has no orientation-preserving actions on the real line. (In algebraic terms, this mean…

Study reduces complexity and uncertainty in human atrial cell models.

This paper proposes using variational autoencoders to model opponents in multi-agent systems.

Removes all bad local minima without adding extra units.

PoWER-BERT speeds up BERT inference by eliminating redundant word-vectors.

GPE algorithm optimizes nonparametric contextual bandits with efficient regret bounds.

We study finite-dimensional integrals in a way that elucidates the mathematical meaning behind the formal manipulations of path integrals occurring in quantum field theory. This involves a proper understanding of how Wick's theorem allows one to evaluate integrals perturbatively, i.e., as a series expansion in a formal…

New RL algorithm reduces policy switching cost to loglog(T) with similar regret.

New model for display advertising with stochastic and adversarial components.