Rethinking CNN’s Generalization to Backdoor Attack from Frequency Domain

Authors: Quanrui Rao, Lin Wang, Wuying Liu

ICLR 2024 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental We conducted experiments on three widely used datasets: CIFAR-10 (Krizhevsky et al. (2009)), Celeba (Liu et al. (2015)) and MNIST (Le Cun et al. (1998)).
Researcher Affiliation Academia 1 Shandong Key Laboratory of Language Resources Development and Application, Ludong University, China 2 School of Information and Electrical Engineering, Ludong University, China 3 Xianda College of Economics and Humanities, Shanghai International Studies University, China
Pseudocode No The paper describes its proposed methods and algorithms in textual form and through equations, but it does not include any explicitly labeled pseudocode or algorithm blocks.
Open Source Code No The paper does not provide any explicit statements about releasing source code for the described methodology, nor does it include links to a code repository.
Open Datasets Yes We conducted experiments on three widely used datasets: CIFAR-10 (Krizhevsky et al. (2009)), Celeba (Liu et al. (2015)) and MNIST (Le Cun et al. (1998)).
Dataset Splits No The paper mentions conducting "validation using the Res Net18...models" but does not provide specific details on the dataset splits for training, validation, or testing for CIFAR-10, Celeba, or MNIST. While it provides train/test splits for Tiny-ImageNet, it does not explicitly mention a validation split for that dataset either.
Hardware Specification No The paper does not provide any specific details regarding the hardware (e.g., GPU models, CPU types, memory) used to conduct the experiments.
Software Dependencies No The paper mentions using the "adam optimizer" and various models like "Res Net18", but it does not provide specific version numbers for any software libraries, frameworks, or programming languages used (e.g., Python, PyTorch, TensorFlow versions).
Experiment Setup Yes During the training phase, we used the adam optimizer, initially using a learning rate of 0.01 and decreasing it by a factor of 10 every 100 training steps.