Robust Architecture

Robust architecture research focuses on designing deep neural networks (DNNs) that are resilient to adversarial attacks and data privacy breaches, aiming to improve both accuracy and robustness. Current research emphasizes the impact of architectural components (e.g., skip connections, residual blocks) on robustness, utilizing techniques like neural architecture search (NAS) and adversarial training to optimize architectures for various attack types (e.g., l∞-norm, semantic attacks). These advancements are crucial for ensuring the reliability and security of DNNs in real-world applications, particularly where data privacy and model integrity are paramount.

Papers

October 25, 2024

Towards Robust Out-of-Distribution Generalization: Data Augmentation and Neural Architecture Search Approaches
Haoyue Bai
Deep Learning Native Robustness Data Augmentation Robust Architecture

September 3, 2024

On the Vulnerability of Skip Connections to Model Inversion Attacks
Jun Hao Koh, Sy-Tuyen Ho, Ngoc-Bao Nguyen, Ngai-man Cheung
Model Inversion Attack Latent Vulnerability Robust Architecture

May 9, 2024

Towards Accurate and Robust Architectures via Neural Architecture Search
Yuwei Ou, Yuqi Feng, Yanan Sun
Adversarial Attack Adversarial Training Neural Architecture Search Adversarial Loss Robust Neural Architecture Robust Architecture

March 19, 2024

Robust NAS under adversarial training: benchmark, theory, and beyond
Yongtao Wu, Fanghui Liu, Carl-Johann Simon-Gabriel, Grigorios G Chrysos, Volkan Cevher
New Benchmark Adversarial Training Neural Architecture Search Theoretical Understanding Deep Learning Theory Robust Architecture Robust Neural Architecture Search

August 30, 2023

Robust Principles: Architectural Design Principles for Adversarially Robust CNNs
ShengYun Peng, Weilin Xu, Cory Cornelius, Matthew Hull, Kevin Li, Rahul Duggal, Mansi Phute, Jason Martin, Duen Horng Chau
Adversarial Training Adversarial Robustness Deep Convolutional Neural Network Robust Framework Robust Architecture Architectural Design

July 18, 2023

An Evaluation of Zero-Cost Proxies -- from Neural Architecture Performance to Model Robustness
Jovita Lukasik, Michael Moeller, Margret Keuper
Native Robustness Model Robustness Zero Cost Proxy Robust Architecture

May 20, 2023

Searching a Compact Architecture for Robust Multi-Exposure Image Fusion
Zhu Liu, Jinyuan Liu, Guanyao Wu, Zihang Chen, Xin Fan, Risheng Liu
Image Pair Deformable Convolution Embracing CompAct Self Alignment Multi Exposure Image Fusion Robust Architecture

May 12, 2023

Efficient Search of Comprehensively Robust Neural Architectures via Multi-fidelity Evaluation
Jialiang Sun, Wen Yao, Tingsong Jiang, Xiaoqian Chen
Multi Fidelity Robust Deep Semantic Adversarial Efficient Search Robust Neural Architecture Robust Architecture Composite Attack

January 8, 2023

RobArch: Designing Robust Architectures against Adversarial Attacks
ShengYun Peng, Weilin Xu, Cory Cornelius, Kevin Li, Rahul Duggal, Duen Horng Chau, Jason Martin
Adversarial Attack Native Robustness Adversarial Training Robust Architecture

December 28, 2022

Differentiable Search of Accurate and Robust Architectures
Yuwei Ou, Xiangning Xie, Shangce Gao, Yanan Sun, Kay Chen Tan, Jiancheng Lv
Adversarial Attack Adversarial Training Differentiable Search Robust Neural Architecture Robust Architecture

December 21, 2022

Revisiting Residual Networks for Adversarial Robustness: An Architectural Perspective
Shihua Huang, Zhichao Lu, Kalyanmoy Deb, Vishnu Naresh Boddeti
Adversarial Robustness Architecture Perspective Residual Block Robust Architecture

January 24, 2022

Hiding Behind Backdoors: Self-Obfuscation Against Generative Models
Siddhartha Datta, Nigel Shadbolt
Generative Model Backdoor Attack Scientific Inference Obfuscation Technique Attack Vector Data Obfuscation Robust Architecture