Accuracy Robustness

Accuracy robustness research focuses on developing machine learning models that maintain high accuracy even when faced with noisy, corrupted, or adversarially perturbed inputs. Current efforts concentrate on improving the accuracy-robustness trade-off through techniques like adversarial training, randomized smoothing, and data augmentation strategies that selectively weight or prune training data based on margin or pixel influence. These advancements are crucial for deploying reliable machine learning systems in high-stakes applications where model accuracy under real-world conditions is paramount, impacting fields ranging from medical diagnosis to autonomous driving.

Papers

November 6, 2024

A Fundamental Accuracy--Robustness Trade-off in Regression and Classification
Sohail Bahmani
Classification Code Adversarial Robustness Novel Regression Adversarial Risk Accuracy Robustness Ridge Function

October 18, 2024

Stochastic Gradient Descent Jittering for Inverse Problems: Alleviating the Accuracy-Robustness Tradeoff
Peimeng Guan, Mark A. Davenport
Native Robustness Stochastic Gradient Descent Inverse Problem Accuracy Robustness

October 17, 2024

Golyadkin's Torment: Doppelgängers and Adversarial Vulnerability
George I. Kamberov
Adversarial Example Meta Learning Generative Adversarial Adversarial Vulnerability Adversarial Imitation Accuracy Robustness Lookalike Model

October 2, 2024

Learning-Augmented Robust Algorithmic Recourse
Kshitij Kayastha, Vasilis Gkatzelis, Shahin Jabbari
Native Robustness Algorithmic Recourse Accuracy Robustness Adversarial Change

July 3, 2024

SPLITZ: Certifiable Robustness via Split Lipschitz Randomized Smoothing
Meiyu Zhong, Ravi Tandon
Adversarial Example Randomized Smoothing Robustness Guarantee Split and Fit Smoothed Classifier Accuracy Robustness Certifiable Robustness

June 2, 2024

Improving Accuracy-robustness Trade-off via Pixel Reweighted Adversarial Training
Jiacheng Zhang, Feng Liu, Dawei Zhou, Jingfeng Zhang, Tongliang Liu
Adversarial Example Adversarial Training Robust Classification Accuracy Robustness Adversarial Reweighting

May 10, 2024

PUMA: margin-based data pruning
Javier Maroto, Pascal Frossard
Native Robustness Adversarial Training Adversarial Perturbation Data Pruning Accuracy Robustness

February 22, 2024

Rethinking Invariance Regularization in Adversarial Training to Improve Robustness-Accuracy Trade-off
Futa Waseda, Ching-Chun Chang, Isao Echizen
Adversarial Example Adversarial Training Adversarial Input Accuracy Robustness Invariance Regularization

December 15, 2023

Closing the Gap: Achieving Better Accuracy-Robustness Tradeoffs against Query-Based Attacks
Pascal Zimmer, Sébastien Andreina, Giorgia Azzurra Marson, Ghassan Karame
Native Robustness Defense Method Low Confidence QuEry Based Attack Accuracy Robustness Perturbation Defense

December 5, 2023

On Optimal Consistency-Robustness Trade-Off for Learning-Augmented Multi-Option Ski Rental
Yongho Shin, Changyeol Lee, Hyung-Chan An
Randomized Algorithm Accuracy Robustness

August 31, 2023

Adversarial Finetuning with Latent Representation Constraint to Mitigate Accuracy-Robustness Tradeoff
Satoshi Suzuki, Shin'ya Yamaguchi, Shoichiro Takeda, Sekitoshi Kanai, Naoki Makishima, Atsushi Ando, Ryo Masumura
Deep Neural Network Adversarial Example Adversarial Training Adversarial Fine Tuning Accuracy Robustness

July 24, 2023

Adaptive Certified Training: Towards Better Accuracy-Robustness Tradeoffs
Zhakshylyk Nurlanov, Frank R. Schmidt, Florian Bernard
Native Robustness Robustness Guarantee Accuracy Robustness Adaptive Training TinyImagenet Benchmark

June 28, 2023

Mitigating Accuracy-Robustness Trade-off via Balanced Multi-Teacher Adversarial Distillation
Shiji Zhao, Xizhe Wang, Xingxing Wei
Adversarial Attack Adversarial Training Accuracy Robustness Multi Teacher Distillation Adversarial Robustness Distillation

May 23, 2023

Expressive Losses for Verified Robustness via Convex Combinations
Alessandro De Palma, Rudy Bunel, Krishnamurthy Dvijotham, M. Pawan Kumar, Robert Stanforth, Alessio Lomuscio
Adversarial Attack Adversarial Training Adversarial Robustness Expressive Power Robustness Verification Accuracy Robustness Convex Combination

January 29, 2023

Improving the Accuracy-Robustness Trade-Off of Classifiers via Adaptive Smoothing
Yatong Bai, Brendon G. Anderson, Aerin Kim, Somayeh Sojoudi
Native Robustness Adversarial Robustness Simple Classifier Robust Classifier Accuracy Robustness Adversarial Region Adaptive Smoothing

November 16, 2022

Differentially Private Optimizers Can Learn Adversarially Robust Models
Yuan Zhang, Zhiqi Bu
Native Robustness Differential Privacy Adversarial Robustness Robust Model Private Optimization Privacy Accuracy Accuracy Robustness

October 10, 2022

Certified Training: Small Boxes are All You Need
Mark Niklas Müller, Franziska Eckert, Marc Fischer, Martin Vechev
Training Data Adversarial Robustness Adversarial Input Object Box Probabilistic Guarantee Accuracy Robustness

June 29, 2022

IBP Regularization for Verified Adversarial Robustness via Branch-and-Bound
Alessandro De Palma, Rudy Bunel, Krishnamurthy Dvijotham, M. Pawan Kumar, Robert Stanforth
Adversarial Attack Gradient Regularization Branch and Bound Accuracy Robustness Adversarial Validation

June 23, 2022

Chasing Convex Bodies and Functions with Black-Box Advice
Nicolas Christianson, Tinashe Handina, Adam Wierman
Machine Learning Algorithm Simple Function Convex Set Convex Function Accuracy Robustness Competitive Algorithm \Delta$ Robustness Machine Learned Advice

June 5, 2022

Vanilla Feature Distillation for Improving the Accuracy-Robustness Trade-Off in Adversarial Training
Guodong Cao, Zhibo Wang, Xiaowei Dong, Zhifei Zhang, Hengchang Guo, Zhan Qin, Kui Ren
Knowledge Distillation Adversarial Example Adversarial Training Adversarial Distillation Accuracy Robustness Adversarial Training Algorithm