arXiv research

Worst-Case Sensitivity measures model sensitivity to uncertainty set size.

Optimizes bond portfolios to avoid worst-case losses.

Study approximates worst-case stock trading under uncertainty, quantifying sensitivity.

RISE learns decisions with sensitive variables, improving worst-case outcomes.

A new method for survival analysis models that ensures fairness without using sensitive demographic data.

A new algorithm avoids worst-case outcomes in risky contexts.

Majorizing measures control sequential complexities for online learning.

Proposes a sensitivity framework to handle limited overlap in causal inference.

A framework identifies worst-case decision points in safety-critical scenarios, improving risk assessment by 10 hours.

This paper concerns the problem of learning control policies for an unknown linear dynamical system to minimize a quadratic cost function. We present a method, based on convex optimization, that accomplishes this task robustly: i.e., we minimize the worst-case cost, accounting for system uncertainty given the observed …

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

The paper shows how to use proxy attributes for fairness in machine learning models with missing sensitive group data.

Deep RL policies are vulnerable to adversarial perturbations, but vanilla training yields more robust policies.

Method identifies shifts leading to large model performance differences.

New method for certified unlearning reduces noise injection.

Machine learning algorithms are increasingly involved in sensitive decision-making process with adversarial implications on individuals. This paper presents mdfa, an approach that identifies the characteristics of the victims of a classifier's discrimination. We measure discrimination as a violation of multi-differenti…

Paper defends sensitive attributes in GNNs from inference attacks.

This paper analyzes risk-sensitive reinforcement learning with Conditional Value-at-Risk (CVaR) for robust Markov Decision Processes.

Paper proposes a new DRL algorithm optimizing Spectral Risk Measures for better risk management.

We study distributed stochastic convex optimization under the delayed gradient model where the server nodes perform parameter updates, while the worker nodes compute stochastic gradients. We discuss, analyze, and experiment with a setup motivated by the behavior of real-world distributed computation networks, where the…

Outliers are ubiquitous in modern data sets. Distance-based techniques are a popular non-parametric approach to outlier detection as they require no prior assumptions on the data generating distribution and are simple to implement. Scaling these techniques to massive data sets without sacrificing accuracy is a challeng…

Paper introduces DOO models to outperform SAA out-of-sample.

A new method for risk-sensitive reinforcement learning using Spectral Risk Measures.

For a stochastic factor model we maximize the long-term growth rate of robust expected power utility with parameter $λ\in(0,1)$. Using duality methods the problem is reformulated as an infinite time horizon, risk-sensitive control problem. Our results characterize the optimal growth rate, an optimal long-term trading s…

A new framework estimates policy value robustly against confounders.

Recently, deep learning has been applied to many security-sensitive applications, such as facial authentication. The existence of adversarial examples hinders such applications. The state-of-the-art result on defense shows that adversarial training can be applied to train a robust model on MNIST against adversarial exa…

This paper studies bounds for the Lipschitz constant of random neural networks.

AdMRL improves meta-reinforcement learning by minimizing worst-case sub-optimality gap.

Distortion (Denneberg 1990) is a well known premium calculation principle for insurance contracts. In this paper, we study sensitivity properties of distortion functionals w.r.t. the assumptions for risk aversion as well as robustness w.r.t. ambiguity of the loss distribution. Ambiguity is measured by the Wasserstein d…

In distributed machine learning, data is dispatched to multiple machines for processing. Motivated by the fact that similar data points often belong to the same or similar classes, and more generally, classification rules of high accuracy tend to be "locally simple but globally complex" (Vapnik & Bottou 1993), we propo…

New bounds assess policy evaluation under unobserved confounders, showing model-based methods are more effective.

We propose to interpret distribution model risk as sensitivity of expected loss to changes in the risk factor distribution, and to measure the distribution model risk of a portfolio by the maximum expected loss over a set of plausible distributions defined in terms of some divergence from an estimated distribution. The…

Proposes a fairness criterion for multi-objective optimization in classification.

Study on adversarial training's impact on deep neural reinforcement learning policies.

Paper introduces Lambda EVaR, a new risk measure.

Study of linear classifiers in infinite imbalance scenarios.

The paper introduces a method to improve adversarial robustness in neural networks using randomized perturbations.

Dropping a tiny fraction of preferences can significantly alter the rankings of top LLMs.

Crowdsourced data used in machine learning services might carry sensitive information about attributes that users do not want to share. Various methods have been proposed to minimize the potential information leakage of sensitive attributes while maximizing the task accuracy. However, little is known about the theory b…

Study risk-sensitive reinforcement learning with entropic risk measures and generative models.

Sharp policy value estimation for contextual bandits with unobserved confounders.

New method uses unlabeled data to improve model robustness across different environments.

Optimal early liquidation strategy reduces financial losses during crises.

We quantify the sensitivity of the Eisenberg-Noe clearing vector to estimation errors in the bilateral liabilities of a financial system in a stylized setting. The interbank liabilities matrix is a crucial input to the computation of the clearing vector. However, in practice central bankers and regulators must often es…

We study the worst-case adaptive optimization problem with budget constraint that is useful for modeling various practical applications in artificial intelligence and machine learning. We investigate the near-optimality of greedy algorithms for this problem with both modular and non-modular cost functions. In both case…

Worst-case risk measures refer to the calculation of the largest value for risk measures when only partial information of the underlying distribution is available. For the popular risk measures such as Value-at-Risk (VaR) and Conditional Value-at-Risk (CVaR), it is now known that their worst-case counterparts can be ev…

Paper improves worst-case regret bounds for RLSVI in reinforcement learning.

Connects robust optimization to conformal prediction for uncertainty sets.