典型文献
Denoised Internal Models:A Brain-inspired Autoencoder Against Adversarial Attacks
文献摘要：
Despite its great success,deep learning severely suffers from robustness;i.e.,deep neural networks are very vulnerable to adversarial attacks,even the simplest ones.Inspired by recent advances in brain science,we propose the denoised internal models(DIM),a novel generative autoencoder-based model to tackle this challenge.Simulating the pipeline in the human brain for visual signal processing,DIM adopts a two-stage approach.In the first stage,DIM uses a denoiser to reduce the noise and the dimensions of inputs,reflecting the information pre-processing in the thalamus.Inspired by the sparse coding of memory-related traces in the primary visual cortex,the second stage produces a set of internal models,one for each category.We evaluate DIM over 42 adversarial attacks,showing that DIM effectively defenses against all the attacks and outperforms the SOTA on the overall robustness on the MNIST(Modified Na-tional Institute of Standards and Technology)dataset.
文献关键词：
中图分类号：
作者姓名：
Kai-Yuan Liu;Xing-Yu Li;Yu-Rui Lai;Hang Su;Jia-Chen Wang;Chun-Xu Guo;Hong Xie;Ji-Song Guan;Yi Zhou
作者机构：
School of Life Sciences and Technology,ShanghaiTech University,Shanghai 201210,China;School of Life Sciences,Tsinghua University,Beijing 100084,China;Shanghai Center for Brain Science and Brain-inspired Technology,Shanghai 201602,China;Institute of Photonic Chips,University of Shanghai for Science and Technology,Shanghai 200093,China;Centre for Artificial-intelligence Nanophotonics,School of Optical-electrical and Computer Engineering,University of Shanghai for Science and Technology,Shanghai 200093,China;National Engineering Laboratory for Brain-inspired Intelligence Technology and Application,School of Information Science and Technology,University of Science and Technology of China,Hefei 230026,China
文献出处：
引用格式：
[1]Kai-Yuan Liu;Xing-Yu Li;Yu-Rui Lai;Hang Su;Jia-Chen Wang;Chun-Xu Guo;Hong Xie;Ji-Song Guan;Yi Zhou-.Denoised Internal Models:A Brain-inspired Autoencoder Against Adversarial Attacks)[J].机器智能研究（英文）,2022(05):456-471
A类：
Denoised,denoiser
B类：
Internal,Models,Brain,inspired,Autoencoder,Against,Adversarial,Attacks,Despite,its,great,success,deep,learning,severely,suffers,from,robustness,neural,networks,are,very,vulnerable,adversarial,attacks,even,simplest,ones,Inspired,by,recent,advances,brain,science,we,propose,denoised,internal,models,DIM,novel,generative,autoencoder,tackle,this,challenge,Simulating,pipeline,human,visual,signal,processing,adopts,stage,approach,first,uses,reduce,dimensions,inputs,reflecting,information,pre,thalamus,sparse,coding,memory,related,traces,primary,cortex,second,produces,each,category,We,evaluate,showing,that,effectively,defenses,against,outperforms,SOTA,overall,MNIST,Modified,Na,tional,Institute,Standards,Technology,dataset
AB值：
0.67908
相似文献
机标中图分类号，由域田数据科技根据网络公开资料自动分析生成，仅供学习研究参考。
