New algorithm reduces reinforcement learning regret by adapting to interaction variability.
problem Existing reinforcement learning methods lack adaptability to interaction variability.
method Developed a variance-adaptive optimal algorithm for MNL function approximation.
result Achieved instance-wise optimal regret bounds, validating efficiency in practice.
New confidence intervals improve treatment effect estimation in randomized experiments.
problem Improving confidence intervals for treatment effects in randomized experiments.
method Systematic exploitation of negative dependence or variance adaptivity.
result Achieved nonasymptotic confidence intervals with the same effective sample size as asymptotic ones.
New algorithm for contextual combinatorial bandits with probabilistic arm triggering.
problem Optimizing decisions in dynamic environments with probabilistic arm availability.
method C^2-UCB-T and VAC^2-UCB algorithms with TPM and VM conditions.
result Achieved improved regret bounds for contextual combinatorial bandits.
Significant improvements in regret analysis for adaptive online learning problems.
problem Exploiting low variance in online learning problems without known variances.
method Novel peeling-based regret analysis leveraging elliptical potential `count` lemma.
result Significant improvements in regret bounds for linear bandits and linear mixture MDPs.
Adaptive linear bandit algorithm with best-of-three-worlds regret bounds.
problem Adaptive to adversarial and stochastic environments with varying sub-optimality gaps and corruption.
method Combines SCRiBLe algorithm with scaled-up sampling and optimistic online learning.
result Achieves best-of-three-worlds regret bounds of O ( T log T ) O(\sqrt{T \log T}) O ( T log T ) for adversarial and O ( log T Δ min + C log T Δ min ) O(\frac{\log T}{Δ_{\min}} + \sqrt{\frac{C \log T}{Δ_{\min}}}) O ( Δ m i n l o g T + Δ m i n C l o g T ) for stochastic environments. RAVEN-UCB addresses non-stationary MAB problems with tighter regret bounds.
problem Non-stationary environments in multi-armed bandits.
method Combines variance-aware adaptation with three innovations: confidence bounds, adaptive control, and recursive updates.
result Achieves tighter regret bounds than UCB1 and UCB-V.
Optimum-statistical collaboration improves black-box optimization efficiency.
problem Improving black-box optimization efficiency through better statistical collaboration.
method Introducing optimum-statistical collaboration framework for hierarchical bandits-based optimization.
result Demonstrated improved regret bounds and better performance in experiments.
EVA adapts LoRA for faster, more efficient fine-tuning.
problem Fast and efficient fine-tuning of large models for specific tasks.
method EVA uses directions capturing most activation variance for initialization, maximizing gradient signal and reducing parameters.
result EVA achieves faster convergence and higher average scores across tasks, reducing parameters.
Novel confidence sets improve linear bandit performance by adapting to unknown noise levels.
problem Adapting to unknown noise levels in sequential decision-making.
method Proposed semi-adaptive and variance-adaptive confidence sets.
result Improved regret bounds and better performance in Bayesian optimization tasks.
The paper offers error bounds for quantized dynamical models.
problem Accuracy of dynamical models from dependent data sequences.
method Developed uniform error bounds for quantized models and imperfect optimization algorithms.
result Unified bounds for slow and fast rates, scaling with model encoding bits.
The ADAM optimizer is exceedingly popular in the deep learning community. Often it works very well, sometimes it doesn't. Why? We interpret ADAM as a combination of two aspects: for each weight, the update direction is determined by the sign of stochastic gradients, whereas the update magnitude is determined by an esti…
The paper proposes a method for distribution-free prediction sets that adapt to unknown temporal changes.
problem Distribution-free prediction sets require reliable calibration data, which is often unavailable in real-world settings with temporal changes.
method The method selects an adaptive window to construct prediction sets, optimizing a bias-variance tradeoff.
result The method provides sharp coverage guarantees and is shown to be adaptive to temporal drift through numerical experiments.
New algorithm reduces best-in-class regret in contextual bandits.
problem Compete with the best policy in a class without model restrictions.
method Proposes an algorithm that updates policies by minimizing a pessimistic objective, including a clipped inverse-propensity estimate and variance penalty.
result Achieves fast best-in-class regret rates, including polylogarithmic rates in the parametric case.
Shampoo achieves higher token efficiency than Muon in language models.
problem Understanding the relationship and relative data efficiency of Shampoo and Muon compared to Adam and Signum.
method Extensive experiments on language models, demonstrating Shampoo's higher efficiency and decomposing its updates.
result Shampoo's benefits are attributed to its application to weight matrices, challenging interpretations based on variance adaptation and whitening.
Improved spectral gap for MwG with adaptive RWM proposals.
problem Improving mixing efficiency of MwG for log-concave distributions.
method Using adaptive RWM proposals tuned to match conditional variances of log-concave target distributions.
result Established a spectral gap lower bound of order O ( 1 / κ d ) \mathcal{O}(1/κd) O ( 1/ κ d ) for MwG. Improved off-policy selection and learning in contextual bandits with better guarantees.
problem Selecting or training a reward-maximizing policy using data from a fixed behavior policy.
method A betting-based confidence bound applied to an inverse propensity weight sequence for off-policy selection, and a freezing condition for off-policy learning.
result The proposed methods achieve significantly improved guarantees over prior work, especially in small-data regimes.
New algorithms improve best-arm identification with varying rewards.
problem Identifying the best arm with varying reward variances in fixed budget.
method Proposed two algorithms: SHVar for known variances, SHAdaVar for unknown variances; uses non-uniform budget allocation.
result Bounding misidentification probabilities for both algorithms.
Paper proves robust M-estimators' coordinates' normality in high dimensions.
problem High-dimensional robust M-estimators' asymptotic normality.
method Develops Stein formulae for high-dimensional random vectors on the sphere.
result Asymptotic normality holds for most coordinates of robust M-estimators with convex penalty.
A federated learning algorithm tackles linear bandits with adversarial actions, achieving optimal regret bounds.
problem Federated linear bandits with finite adversarial action sets.
method FedSupLinUCB algorithm, extending SupLinUCB and OFUL principles.
result Achieves a total regret of i l d e O ( d T ) ilde{O}(\sqrt{d T}) i l d e O ( d T ) , matching minimax lower bound and being order-optimal. New algorithm reduces regret for linear bandits with unknown noise variance.
problem Finding optimal actions in linear bandits with varying noise variance.
method Adaptive algorithm with Freedman-type concentration inequality and multi-layer structure.
result Achieves i l d e O ( d ∑ k = 1 K σ k 2 + d ) ilde{O}(d \sqrt{\sum_{k = 1}^K σ_k^2} + d) i l d e O ( d ∑ k = 1 K σ k 2 + d ) regret for linear bandits.