Robust Neural Architecture

Robust neural architecture research focuses on designing and discovering deep learning models that are resilient to various forms of noise and adversarial attacks, maintaining high accuracy under challenging conditions. Current efforts utilize neural architecture search (NAS) algorithms, often incorporating multi-objective optimization and multi-fidelity evaluations to efficiently explore the vast design space and find architectures robust against diverse attacks (e.g., l<sub>p</sub> norm attacks, semantic attacks). This field is crucial for deploying reliable AI systems in real-world applications where data imperfections and malicious manipulations are common, impacting the safety and security of numerous domains.

Papers

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 Architecture Robust Neural Architecture

May 16, 2023

Noise robust neural network architecture
Xiong Yunuo, Xiong Hongwei
Neural Network Deep Neural Network Industrial Disturbing Noise Neural Network Architecture Noisy Image Robust Neural 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 Architecture Robust Neural Architecture Composite Attack

February 15, 2023

XploreNAS: Explore Adversarially Robust & Hardware-efficient Neural Architectures for Non-ideal Xbars
Abhiroop Bhattacharjee, Abhishek Moitra, Priyadarshini Panda
Adversarial Attack Adversarial Training Adversarial Robustness Crossbar Array Memristive Crossbar Robust Neural Architecture

January 19, 2023

RNAS-CL: Robust Neural Architecture Search by Cross-Layer Knowledge Distillation
Utkarsh Nath, Yancheng Wang, Yingzhen Yang
Adversarial Attack Neural Architecture Search Robust Neural Architecture Robust Neural Architecture Search

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

October 18, 2022

Less is More: A Lightweight and Robust Neural Architecture for Discourse Parsing
Ming Li, Ruihong Huang
Pre Trained Language Model Neural Language Model Discourse Parsing Robust Neural Architecture Discourse Parser

July 12, 2022

Bi-fidelity Evolutionary Multiobjective Search for Adversarially Robust Deep Neural Architectures
Jia Liu, Ran Cheng, Yaochu Jin
Adversarial Attack Adversarial Training Adversarial Robustness Multiobjective Optimization Multi Objective Neural Architecture Search Evolutionary Multiobjective Robust Neural Architecture

June 7, 2022

Can CNNs Be More Robust Than Transformers?
Zeyu Wang, Yutong Bai, Yuyin Zhou, Cihang Xie
Convolutional Neural Network Vision Transformer Transformer Megatron Decepticons CNN Model Robust Neural Architecture

April 21, 2022

BTranspose: Bottleneck Transformers for Human Pose Estimation with Self-Supervised Pre-Training
Kaushik Balakrishnan, Devesh Upadhyay
Human Pose Estimation Self Supervised Pre Training Multi Head Self Attention 2 Dimensional Human Pose Estimation Robust Neural Architecture Bottleneck Transformer Transformer Encoder Layer

March 7, 2022

$A^{3}D$: A Platform of Searching for Robust Neural Architectures and Efficient Adversarial Attacks
Jialiang Sun, Wen Yao, Tingsong Jiang, Chao Li, Xiaoqian Chen
Adversarial Attack Search Query Adversarial Noise Diverse Platform Robust Neural Architecture Efficient Adversarial Attack

Robust Neural Architecture

Papers

Towards Accurate and Robust Architectures via Neural Architecture Search

Noise robust neural network architecture

Efficient Search of Comprehensively Robust Neural Architectures via Multi-fidelity Evaluation

XploreNAS: Explore Adversarially Robust & Hardware-efficient Neural Architectures for Non-ideal Xbars

RNAS-CL: Robust Neural Architecture Search by Cross-Layer Knowledge Distillation

Differentiable Search of Accurate and Robust Architectures

Less is More: A Lightweight and Robust Neural Architecture for Discourse Parsing

Bi-fidelity Evolutionary Multiobjective Search for Adversarially Robust Deep Neural Architectures

Can CNNs Be More Robust Than Transformers?

BTranspose: Bottleneck Transformers for Human Pose Estimation with Self-Supervised Pre-Training

$A^{3}D$: A Platform of Searching for Robust Neural Architectures and Efficient Adversarial Attacks