Distributionally Robust Optimization

Distributionally robust optimization (DRO) is a framework for training machine learning models that are robust to uncertainty in the data distribution, aiming to minimize the worst-case risk across a set of possible distributions. Current research focuses on developing efficient algorithms for various DRO formulations, including those based on Wasserstein distance, f-divergences, and other metrics, often incorporating techniques like variance reduction and primal-dual methods to improve scalability and convergence. DRO's significance lies in its ability to improve the generalization and fairness of models, particularly in applications with limited data, noisy labels, or significant distributional shifts, leading to more reliable and equitable decision-making in diverse fields.

Papers

February 8, 2024

An Inexact Halpern Iteration with Application to Distributionally Robust Optimization
Ling Liang, Kim-Chuan Toh, Jia-Jie Zhu
Application Proficiency Distributionally Robust Optimization Convex Concave Robust Optimization Problem Convergence Property Monotone Inclusion Halpern Iteration

February 2, 2024

A Distributionally Robust Optimisation Approach to Fair Credit Scoring
Pablo Casas, Christophe Mues, Huan Yu
Fairness Metric Distributionally Robust Optimization Credit Scoring Fair Credit

January 28, 2024

Bayesian Nonparametrics Meets Data-Driven Distributionally Robust Optimization
Nicola Bariletto, Nhat Ho
Empirical Risk Minimization Robust Optimization Distributionally Robust Optimization Bayesian Nonparametric Distributional Uncertainty Data Driven Risk

December 20, 2023

BSL: Understanding and Improving Softmax Loss for Recommendation
Junkang Wu, Jiawei Chen, Jiancan Wu, Wentao Shi, Jizhi Zhang, Xiang Wang
Loss Function Human Understanding Effective Recommendation Recommendation Model Distributionally Robust Optimization Softmax Loss

December 3, 2023

Anomaly Detection Under Uncertainty Using Distributionally Robust Optimization Approach
Amir Hossein Noormohammadia, Seyed Ali MirHassania, Farnaz Hooshmand Khaligh
Anomaly Detection High Uncertainty Anticipation Support Vector Machine Distributionally Robust Optimization One Class Support Vector Machine

November 8, 2023

Geometry-Calibrated DRO: Combating Over-Pessimism with Free Energy Implications
Jiashuo Liu, Jiayun Wu, Tianyu Wang, Hao Zou, Bo Li, Peng Cui
Parameter Estimation Distributionally Robust Optimization Free Energy Error Bound Minimax Optimization Entropic Risk

November 6, 2023

DRAUC: An Instance-wise Distributionally Robust AUC Optimization Framework
Siran Dai, Qianqian Xu, Zhiyong Yang, Xiaochun Cao, Qingming Huang
Distributionally Robust Optimization Area Under the ROC Curve Instance Level Loss

November 3, 2023

Obtaining Explainable Classification Models using Distributionally Robust Optimization
Sanjeeb Dash, Soumyadip Ghosh, Joao Goncalves, Mark S. Squillante
High Explainability Generalization Performance Model Explanation Distributionally Robust Optimization Explainable Classification

October 30, 2023

Flow-based Distributionally Robust Optimization
Chen Xu, Jonghyeok Lee, Xiuyuan Cheng, Yao Xie
Robust Optimization Worst Case Distributionally Robust Optimization Robust Algorithm Transport Map Wasserstein Uncertainty

October 21, 2023

Distributionally Robust Optimization with Bias and Variance Reduction
Ronak Mehta, Vincent Roulet, Krishna Pillutla, Zaid Harchaoui
Absolute Stance Bias Stochastic Gradient Robust Optimization Risk Sensitive Variance Reduction Distributionally Robust Optimization F$ Divergence Spectral Risk

October 17, 2023

Understanding Contrastive Learning via Distributionally Robust Optimization
Junkang Wu, Jiawei Chen, Jiancan Wu, Wentao Shi, Xiang Wang, Xiangnan He
Contrastive Learning Native Robustness Mutual Information Robust Optimization Distributionally Robust Optimization

September 29, 2023

Out-Of-Domain Unlabeled Data Improves Generalization
Amir Hossein Saberi, Amir Najafi, Alireza Heidari, Mohammad Hosein Movasaghinia, Abolfazl Motahari, Babak H. Khalaj
Strong Generalization Distributionally Robust Optimization Domain Sample Robust Loss Function

September 15, 2023

Wasserstein Distributionally Robust Policy Evaluation and Learning for Contextual Bandits
Yi Shen, Pan Xu, Michael M. Zavlanos
LeArning Abstract Gradient Descent Optimal Policy Contextual Bandit Policy Evaluation Robust Optimization Distributionally Robust Optimization

September 12, 2023

Distributionally Robust Transfer Learning
Xin Xiong, Zijian Guo, Tianxi Cai
Transfer Learning Adversarial Loss Distributionally Robust Optimization Distributionally Robust Distributional Robustness

August 12, 2023

Distributionally Robust Optimization and Invariant Representation Learning for Addressing Subgroup Underrepresentation: Mechanisms and Limitations
Nilesh Kumar, Ruby Shrestha, Zhiyuan Li, Linwei Wang
Fundamental Limitation Invariant Representation Functional Mechanism Distributionally Robust Optimization Underrepresented Group Subgroup Performance

July 25, 2023

Federated Distributionally Robust Optimization with Non-Convex Objectives: Algorithm and Analysis
Yang Jiao, Kai Yang, Dongjin Song
Native Robustness General Analysis Practical Algorithm Robust Optimization Distributionally Robust Optimization Non Convex Objective Robustness Trade

June 24, 2023

Smoothed $f$-Divergence Distributionally Robust Optimization
Zhenyuan Liu, Bart P. G. Van Parys, Henry Lam
Average Approximation Robust Optimization Distributionally Robust Optimization Superior Optimizer Data Driven Optimization Smoothed Analysis

June 15, 2023

Stochastic Re-weighted Gradient Descent via Distributionally Robust Optimization
Ramnath Kumar, Kushal Majmundar, Dheeraj Nagaraj, Arun Sai Suggala
Deep Neural Network Robust Optimization Distributionally Robust Optimization Dynamic Reweighting

June 5, 2023

Nonlinear Distributionally Robust Optimization
Mohammed Rayyan Sheriff, Peyman Mohajerin Esfahani
Robust Optimization Distributionally Robust Optimization Optimization Solver Portfolio Optimization Problem

May 22, 2023

Achieving the Minimax Optimal Sample Complexity of Offline Reinforcement Learning: A DRO-Based Approach
Yue Wang, Jinjun Xiong, Shaofeng Zou
Offline Reinforcement Learning Sample Complexity Robust Optimization Distributionally Robust Optimization Transition Kernel