Perturbation Robustness

Perturbation robustness focuses on developing models and algorithms that maintain reliable performance despite noisy or corrupted inputs, a crucial challenge across diverse fields like machine learning and robotics. Current research emphasizes improving robustness through techniques such as control barrier functions (in robotics), adversarial training and hypernetworks (in image recognition and other machine learning tasks), and noise-alignment pre-training (in natural language processing). These advancements are vital for building trustworthy and reliable systems in real-world applications where imperfect data is the norm, impacting fields ranging from autonomous systems to medical diagnosis.

Papers

September 23, 2024

Maintaining Strong $r$-Robustness in Reconfigurable Multi-Robot Networks using Control Barrier Functions
Haejoon Lee, Dimitra Panagou
Native Robustness Control Barrier Function Estimated Team Strength Leader Follower Robot Network Perturbation Robustness

August 2, 2024

Interpreting Global Perturbation Robustness of Image Models using Axiomatic Spectral Importance Decomposition
Róisín Luo, James McDermott, Colm O'Riordan
Image Modeling Mechanistic Interpretability Global Interpretability Global Robustness Spectral Decomposition Perturbation Robustness

April 11, 2024

Efficient Duple Perturbation Robustness in Low-rank MDPs
Yang Hu, Haitong Ma, Bo Dai, Na Li
Native Robustness Low Rank Robustness Evaluation Perturbation Robustness

February 22, 2024

Noise-BERT: A Unified Perturbation-Robust Framework with Noise Alignment Pre-training for Noisy Slot Filling Task
Jinxu Zhao, Guanting Dong, Yueyan Qiu, Tingfeng Hui, Xiaoshuai Song, Daichi Guo, Weiran Xu
Slot Filling Input Perturbation Masked Prediction Perturbation Robustness Noise BERT

November 15, 2023

Multi-Set Inoculation: Assessing Model Robustness Across Multiple Challenge Sets
Vatsal Gupta, Pranshu Pandya, Tushar Kataria, Vivek Gupta, Dan Roth
Large Language Model Language Model Pre Trained Model Model Robustness Input Perturbation Perturbation Robustness Multiset Neuron

September 28, 2023

Parameter-Saving Adversarial Training: Reinforcing Multi-Perturbation Robustness via Hypernetworks
Huihui Gong, Minjing Dong, Siqi Ma, Seyit Camtepe, Surya Nepal, Chang Xu
Adversarial Attack Adversarial Training Adversarial Perturbation Attention Hypernetworks Perturbation Robustness

May 1, 2023

Towards the Flatter Landscape and Better Generalization in Federated Learning under Client-level Differential Privacy
Yifan Shi, Kang Wei, Li Shen, Yingqi Liu, Xueqian Wang, Bo Yuan, Dacheng Tao
Better Generalization Differentially Private Differentially Private Federated Learning Perturbation Robustness Client Level Differential Privacy Return Landscape

March 15, 2023

PR-MCS: Perturbation Robust Metric for MultiLingual Image Captioning
Yongil Kim, Yerin Hwang, Hyeongu Yun, Seunghyun Yoon, Trung Bui, Kyomin Jung
Image Captioning Level Perturbation Perturbation Robustness

February 11, 2022

Robust estimation algorithms don't need to know the corruption level
Ayush Jain, Alon Orlitsky, Vaishakh Ravindrakumar
Robust Estimation Corruption Level Perturbation Robustness Optimal Accuracy

Perturbation Robustness

Papers

Maintaining Strong $r$-Robustness in Reconfigurable Multi-Robot Networks using Control Barrier Functions

Interpreting Global Perturbation Robustness of Image Models using Axiomatic Spectral Importance Decomposition

Efficient Duple Perturbation Robustness in Low-rank MDPs

Noise-BERT: A Unified Perturbation-Robust Framework with Noise Alignment Pre-training for Noisy Slot Filling Task

Multi-Set Inoculation: Assessing Model Robustness Across Multiple Challenge Sets

Parameter-Saving Adversarial Training: Reinforcing Multi-Perturbation Robustness via Hypernetworks

Towards the Flatter Landscape and Better Generalization in Federated Learning under Client-level Differential Privacy

PR-MCS: Perturbation Robust Metric for MultiLingual Image Captioning

Robust estimation algorithms don't need to know the corruption level