arXiv research

The paper develops a new algorithm for constructing minimax estimators using online learning techniques.

Many tasks in modern machine learning can be formulated as finding equilibria in \emph{sequential} games. In particular, two-player zero-sum sequential games, also known as minimax optimization, have received growing interest. It is tempting to apply gradient descent to solve minimax optimization given its popularity a…

Adversarial meta-learning computes Gamma-minimax estimators for vague prior knowledge.

State-of-the-art adversarial attacks are aimed at neural network classifiers. By default, neural networks use gradient descent to minimize their loss function. The gradient of a classifier's loss function is used by gradient-based adversarial attacks to generate adversarially perturbed images. We pose the question whet…

New bounds on minimax regret for sequential probability assignment using logarithmic loss.

Classical scaling is shown to be optimal under various noisy conditions.

MRCpy implements minimax risk classifiers with performance guarantees and distribution shift adaptability.

Given a task of predicting $Y$ from $X$, a loss function $L$, and a set of probability distributions $Γ$ on $(X,Y)$, what is the optimal decision rule minimizing the worst-case expected loss over $Γ$? In this paper, we address this question by introducing a generalization of the principle of maximum entropy. Applying t…

We prove a new minimax theorem connecting the worst-case Bayesian regret and minimax regret under partial monitoring with no assumptions on the space of signals or decisions of the adversary. We then generalise the information-theoretic tools of Russo and Van Roy (2016) for proving Bayesian regret bounds and combine th…

Study reveals differences in label shift problem difficulty in supervised vs. unsupervised settings.

Study minimax linear regression under quantile risk, improving existing bounds and providing new results.

New RL algorithm minimizes distributional learning error.

Paper defends machine learning models from adversarial attacks using GLRT.

MOTS improves Thompson sampling to match minimax bounds for bandit problems.

It is generally believed that ensemble approaches, which combine multiple algorithms or models, can outperform any single algorithm at machine learning tasks, such as prediction. In this paper, we propose Bayesian convex and linear aggregation approaches motivated by regression applications. We show that the proposed a…

Study confirms optimal minimax rate for nonlocal interaction kernel estimation.

Private two-sample tests under LDP achieve minimax rates for multinomial and continuous data.

New approach transfers rewards learned in one environment to reinforcement learning in a new environment.

GL-LowPopArt improves minimax-optimal estimation for trace regression.

Here we propose a general theoretical method for analyzing the risk bound in the presence of adversaries. Specifically, we try to fit the adversarial learning problem into the minimax framework. We first show that the original adversarial learning problem can be reduced to a minimax statistical learning problem by intr…

ConquerNet smooths quantile regression for deep learning with minimax guarantees.

Develops GLRT for defending against adversarial attacks in hypothesis testing.

This work optimizes identifying good arms in nonparametric multi-armed bandits.

New SGDA method speeds up nonconvex minimax optimization.

Measuring the similarity of two files is an important task in malware analysis, with fuzzy hash functions being a popular approach. Traditional fuzzy hash functions are data agnostic: they do not learn from a particular dataset how to determine similarity; their behavior is fixed across all datasets. In this paper, we …

We study online convex optimization under stochastic sub-gradient observation faults, where we introduce adaptive algorithms with minimax optimal regret guarantees. We specifically study scenarios where our sub-gradient observations can be noisy or even completely missing in a stochastic manner. To this end, we propose…

Proves minimax sample complexity for turn-based stochastic games.

Neural networks minimize error with shallow ReLU models for function estimation.

MOSAIC detects change points in dynamic networks with low-rank and sparse changes.

The paper analyzes reinforcement learning methods for estimating weights and quality functions with fast convergence rates.

Two new algorithms solve nonconvex-strongly concave problems efficiently.

New methods solve complex optimization problems without strong convexity assumptions.

Score attack method provides a lower bound on privacy-constrained minimax risk.

The paper tackles robust policy learning from multiple data sources.

Paper develops a method for estimating PFLM with minimized rates in high dimensions.

New algorithm solves nonconvex-convex minimax problems efficiently.

This work provides a simplified proof of the statistical minimax optimality of (iterate averaged) stochastic gradient descent (SGD), for the special case of least squares. This result is obtained by analyzing SGD as a stochastic process and by sharply characterizing the stationary covariance matrix of this process. The…

Paper develops a minimax optimal test for goodness-of-fit using kernel Stein discrepancy.

Unified analysis of efficient local training methods for distributed variational inequalities.

Incremental clustering approaches have been proposed for handling large data when given data set is too large to be stored. The key idea of these approaches is to find representatives to represent each cluster in each data chunk and final data analysis is carried out based on those identified representatives from all t…

Study stabilizes adversarial training in neural networks over infinite-dimensional spaces.

A novel decentralized algorithm improves minimax optimization in federated learning.

EQO uses a simple bonus term for efficient exploration in tabular RL.

Paper proposes a federated learning framework for relative fairness.

Efficient strategies for online learning against bandit algorithms solve minimax problems.

We investigate the use of Minimax distances to extract in a nonparametric way the features that capture the unknown underlying patterns and structures in the data. We develop a general-purpose and computationally efficient framework to employ Minimax distances with many machine learning methods that perform on numerica…

New statistical framework for coresets in density estimation.

Preconditioned non-convex gradient descent improves noisy matrix estimation.