Detecting Adversarial Examples from Sensitivity Inconsistency of Spatial-Transform Domain

Authors: Jinyu Tian, Jiantao Zhou, Yuanman Li, Jia Duan9877-9885

AAAI 2021 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental Intensive experimental results on Res Net and VGG validate the superiority of the proposed SID.
Researcher Affiliation Academia 1State Key Laboratory of Internet of Things for Smart City, Department of Computer and Information Science, University of Macau 2Guangdong Key Laboratory of Intelligent Information Processing, College of Electronics and Information Engineering, Shenzhen University
Pseudocode Yes Algorithm 1: Training procedure of SID
Open Source Code No The paper does not provide any explicit statement about releasing source code or links to a code repository for the described methodology.
Open Datasets Yes We consider two network structures VGG19 with batch normalization (Simonyan and Zisserman 2014) and Res Net34 (He et al. 2016) on two datasets CIFAR10 (Krizhevsky and Hinton 2009) and SVHN (Netzer et al. 2011).
Dataset Splits No The paper mentions using test images for evaluation and randomly selecting examples from the test set, but it does not specify explicit training/validation/test dataset splits (e.g., percentages or counts for each split, or references to predefined standard splits for reproducibility).
Hardware Specification No The paper does not provide specific hardware details (e.g., GPU/CPU models, memory) used for running the experiments.
Software Dependencies No The paper states that 'In the WAWT, the wavelet transform sym17 is adopted (Lee et al. 2019).' The citation refers to 'Pywavelets: A Python Package for Wavelet Analysis,' but no specific version for Pywavelets or other core software (e.g., Python, PyTorch, TensorFlow) is provided.
Experiment Setup No The paper states that 'The perturbation magnitudes η of all the AEs can be found in the supplementary file,' implying some experimental setup details are provided there, but these details are not present in the main text of the paper. No other specific hyperparameters or detailed training configurations are explicitly listed within the main content.