Introduction to NIPS 2017 Competition Track

Escalera, Sérgio; Weimer, Markus; Burtsev, Mikhail; Malykh, Valentin; Logacheva, Varvara; Lowe, Ryan; Serban, Iulian Vlad; Bengio, Yoshua; Rudnicky, Alexander I.; Black, Alan W.; Prabhumoye, Shrimai; Kidziński, Łukasz; Mohanty, Sharada P.; Ong, Carmichael; Hicks, Jennifer L.; Levine, Sergey; Salathé, Marcel; Delp, Scott L.; Huerga, Iker; Григоренко, А. Н.; Thorbergsson, Leifur; Das, Anasuya; Nemitz, Kyla; Sandker, Jenna; King, Stephen J.; Ecker, Alexander; Gatys, Leon A.; Bethge, Matthias; Boyd-Graber, Jordan; Feng, Shi; Rodríguez, Pedro; Iyyer, Mohit; He, He; DauméIII, Hal; McGregor, Sean; Banifatemi, Amir; Kurakin, Alexey; Goodfellow, Ian J.; Bengio, Samy

doi:10.1007/978-3-319-94042-7_1

Cited by 1 publication

(1 citation statement)

References 3 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Data set : Consistent with TAP, we analyze the transferability of our method on an ImageNet‐compatible data set 57 in this paper, which consists of 1000 images that are not used in the original ImageNet data set. Besides, to avoid overfitting, we also randomly select 1000 images from the ILSVRC2012 validation set to perform hyperparameters selection during experiments.…”

Section: Methodsmentioning

confidence: 99%

Attention‐guided transformation‐invariant attack for black‐box adversarial examples

Zhu

Dai

et al. 2022

Int J of Intelligent Sys

View full text Add to dashboard Cite

With the development of media convergence, information acquisition is no longer limited to traditional media, such as newspapers and televisions, but more from digital media on the Internet, where media contents should be under supervision by platforms. At present, the media content analysis technology of Internet platforms relies on deep neural networks (DNNs). However, DNNs show vulnerability to adversarial examples, which results in security risks. Therefore, it is necessary to adequately study the internal mechanism of adversarial examples to build more effective supervision models. When coming to practical applications, supervision models are mostly faced with black‐box attacks, where cross‐model transferability of adversarial examples has attracted increasing attention. In this paper, to improve the transferability of adversarial examples, we propose an attention‐guided transformation‐invariant adversarial attack method, which incorporates an attention mechanism to disrupt the most distinctive features and simultaneously ensures adversarial attack invariance under different transformations. Specifically, we dynamically weight the latent features according to an attention mechanism and disrupt them accordingly. Meanwhile, considering the lack of semantics in low‐level features, high‐level semantics are introduced as spatial guidance to make low‐level feature perturbations concentrate on the most discriminative regions. Moreover, since the attention heatmaps may vary significantly across different models, a transformation‐invariant aggregated attack strategy is proposed to alleviate overfitting to the proxy model attention. Comprehensive experimental results show that the proposed method can significantly improve the transferability of adversarial examples.

show abstract

Section: Methodsmentioning

confidence: 99%