Double Robustness

Double robustness aims to create models that remain reliable even when underlying assumptions are partially incorrect, a crucial goal in various fields facing noisy or incomplete data. Current research focuses on extending this concept beyond single-source errors, addressing multiple simultaneous attacks (e.g., in adversarial machine learning) or handling data missing not at random (e.g., in recommender systems). This involves developing algorithms like dually robust actor-critic methods and stabilized doubly robust estimators, improving model stability and accuracy under uncertainty. The impact of this work spans diverse applications, from enhancing the safety and reliability of AI systems to improving the accuracy of causal inference in complex settings.

Papers

May 2, 2024

Position: Towards Resilience Against Adversarial Examples
Sihui Dai, Chong Xiang, Tong Wu, Prateek Mittal
Native Robustness Robust Continual Learning Trigger Attack Multiple Attack Attack Robustness Double Robustness

September 13, 2023

Safe Reinforcement Learning with Dual Robustness
Zeyang Li, Chuxiong Hu, Yunan Wang, Yujie Yang, Shengbo Eben Li
Reinforcement Learning Deep Reinforcement Learning Safe Reinforcement Learning Robust Reinforcement Learning Policy Iteration Double Robustness

June 27, 2023

Cooperation or Competition: Avoiding Player Domination for Multi-Target Robustness via Adaptive Budgets
Yimu Wang, Dinghuai Zhang, Yihan Wu, Heng Huang, Hongyang Zhang
Adversarial Attack Long Term Cooperation Competition Platform Adversary Agent Monotone Adversary Robust Network Double Robustness Matrix Domination

September 19, 2022

Multilevel Robustness for 2D Vector Field Feature Tracking, Selection, and Comparison
Lin Yan, Paul Aaron Ullrich, Luke P. Van Roekel, Bei Wang, Hanqi Guo
Native Robustness Consistent Comparison View Selection Vector Field Point Tracking Double Robustness

July 21, 2022

AugRmixAT: A Data Processing and Training Method for Improving Multiple Robustness and Generalization Performance
Xiaoliang Liu, Furao Shen, Jian Zhao, Changhai Nie
Native Robustness Adversarial Robustness Generalization Performance Training Method Data Processing Double Robustness

May 10, 2022

StableDR: Stabilized Doubly Robust Learning for Recommendation on Data Missing Not at Random
Haoxuan Li, Chunyuan Zheng, Peng Wu
Raw Data Recommender System Effective Recommendation Robust Learning Imputation Algorithm Random Participation Double Robustness

March 24, 2022

A Rationale-Centric Framework for Human-in-the-loop Machine Learning
Jinghui Lu, Linyi Yang, Brian Mac Namee, Yue Zhang
Human in the Loop Rationale Generation Distribution Performance Double Robustness

November 12, 2021

Dynamic treatment effects: high-dimensional inference under model misspecification
Yuqian Zhang, Weijie Ji, Jelena Bradic
High Dimensional Robust Estimation Model Misspecification Double Robustness Dynamic Treatment Effect

Double Robustness

Papers

Position: Towards Resilience Against Adversarial Examples

Safe Reinforcement Learning with Dual Robustness

Cooperation or Competition: Avoiding Player Domination for Multi-Target Robustness via Adaptive Budgets

Multilevel Robustness for 2D Vector Field Feature Tracking, Selection, and Comparison

AugRmixAT: A Data Processing and Training Method for Improving Multiple Robustness and Generalization Performance

StableDR: Stabilized Doubly Robust Learning for Recommendation on Data Missing Not at Random

A Rationale-Centric Framework for Human-in-the-loop Machine Learning

Dynamic treatment effects: high-dimensional inference under model misspecification