arXiv research

An algorithm learns from multiple models to match an oracle's risk.

New data-dependent priors improve PAC-Bayes bounds.

New oracles improve stochastic optimization with noisy or biased measurements.

Optimizes bilevel empirical risk minimization with improved oracle calls.

New bound improves on weighted majority vote risk estimation.

Study on gradient complexity of private optimization with private oracles.

Improved algorithm finds second-order stationary points in non-convex optimization.

Develops new methods to estimate treatment effects in survival data with competing risks.

Paper analyzes sparse aggregation in GLMs with Kullback-Leibler risk bounds.

We investigate properties of estimators obtained by minimization of U-processes with the Lasso penalty in high-dimensional settings. Our attention is focused on the ranking problem that is popular in machine learning. It is related to guessing the ordering between objects on the basis of their observed predictors. We p…

New algorithm predicts piecewise regular functions online.

EP-learning framework improves causal contrast estimation efficiency.

Propose an XMSE-aware mixed estimator for EB that interpolates between ML and EB shrinkage.

New stability bounds for GD in overparameterised shallow nets without NTK assumptions.

Aggregated hold-out (Agghoo) is a method which averages learning rules selected by hold-out (that is, cross-validation with a single split). We provide the first theoretical guarantees on Agghoo, ensuring that it can be used safely: Agghoo performs at worst like the hold-out when the risk is convex. The same holds true…

The paper analyzes variational autoencoders for state space models with risk bounds.

Derives metrics for DeFi vaults, addressing credit risk.

Piecewise polynomial interpolation-based gradient descent reduces oracle complexity for smooth loss functions.

The paper analyzes prediction error in nonstationary settings using weighted risk minimization.

Two important goals of high-dimensional modeling are prediction and variable selection. In this article, we consider regularization with combined $L_1$ and concave penalties, and study the sampling properties of the global optimum of the suggested method in ultra-high dimensional settings. The $L_1$-penalty provides th…

Discussion of ``2004 IMS Medallion Lecture: Local Rademacher complexities and oracle inequalities in risk minimization'' by V. Koltchinskii [arXiv:0708.0083]

Discussion of ``2004 IMS Medallion Lecture: Local Rademacher complexities and oracle inequalities in risk minimization'' by V. Koltchinskii [arXiv:0708.0083]

Discussion of ``2004 IMS Medallion Lecture: Local Rademacher complexities and oracle inequalities in risk minimization'' by V. Koltchinskii [arXiv:0708.0083]

Discussion of ``2004 IMS Medallion Lecture: Local Rademacher complexities and oracle inequalities in risk minimization'' by V. Koltchinskii [arXiv:0708.0083]

Discussion of "2004 IMS Medallion Lecture: Local Rademacher complexities and oracle inequalities in risk minimization" by V. Koltchinskii [arXiv:0708.0083]

Study on tradeoffs between mistakes and ERM oracle calls in online and transductive learning.

New method improves solving combinatorial optimization problems with smoothed policies.

We presented Bayesian portfolio selection strategy, via the $k$ factor asset pricing model. If the market is information efficient, the proposed strategy will mimic the market; otherwise, the strategy will outperform the market. The strategy depends on the selection of a portfolio via Bayesian multiple testing methodol…

New algorithm reduces ERM problem size while maintaining accuracy.

AGCA approximates angular variation on the unit sphere, reducing extremal dependence problems to eigenanalysis.

The paper tackles deep learning from dependent data, achieving optimal performance.

New lower bounds for bilevel optimization with first-order oracles.

New algorithm reduces sample complexity for multi-distribution learning.

Quantum method speeds up risk estimation for insurance tail risks.

B-Learner provides bounds on CATE under hidden confounding risks.

Paper analyzes risk bounds for in-context learning in multiclass classification.

The paper tackles fair set-valued classification under demographic parity constraints.

In this paper we revisit the risk bounds of the lasso estimator in the context of transductive and semi-supervised learning. In other terms, the setting under consideration is that of regression with random design under partial labeling. The main goal is to obtain user-friendly bounds on the off-sample prediction risk.…

New algorithm learns efficiently with a simple 'yes/no' oracle.

Proposed by Donoho (1997), Dyadic CART is a nonparametric regression method which computes a globally optimal dyadic decision tree and fits piecewise constant functions in two dimensions. In this article we define and study Dyadic CART and a closely related estimator, namely Optimal Regression Tree (ORT), in the contex…

New RL algorithm maximizes CVaR in low-rank MDPs with provable efficiency.

New algorithms reduce reinforcement learning regret in factored MDPs.

Quantum method improves CVaR evaluation under correlated fields.

LOO prediction method improves generalization guarantees for arbitrary datasets.

We study the fundamental tradeoffs between computational tractability and statistical accuracy for a general family of hypothesis testing problems with combinatorial structures. Based upon an oracle model of computation, which captures the interactions between algorithms and data, we establish a general lower bound tha…

Proposes a nonparametric model for dynamic team rankings.

New method for estimating and optimizing MDPs without stationarity.

New algorithm reduces regret bounds for Bayesian optimization with unknown hyperparameters.