Spectral Signatures in Backdoor Attacks

Authors: Brandon Tran, Jerry Li, Aleksander Madry

NeurIPS 2018 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental We demonstrate the efficacy of these signatures in detecting and removing poisoned examples on real image sets and state of the art neural network architectures.
Researcher Affiliation Collaboration Brandon Tran EECS MIT Cambridge, MA 02139 btran@mit.edu Jerry Li Simons Institute Berkeley, CA 94709 jerryzli@berkeley.edu Aleksander M adry EECS MIT madry@mit.edu ... J.L. was supported by ... and an intern at Google Brain. ... A.M. was supported in part by ... a Google Research Award...
Pseudocode Yes More detailed pseudocode is provided in Algorithm 1.
Open Source Code No The paper does not contain an explicit statement about releasing its source code or a link to a code repository.
Open Datasets Yes On CIFAR-10, which contains 5000 images for each of 10 labels... We study backdoor poisoning attacks on the CIFAR10 [19] dataset...
Dataset Splits No The paper mentions training and test sets (e.g., 'original test set', 'backdoored test set' for CIFAR-10) but does not specify validation splits or detailed percentages for data partitioning.
Hardware Specification No The paper does not explicitly describe the specific hardware (e.g., GPU/CPU models, memory) used to run its experiments.
Software Dependencies No The paper does not provide specific software dependencies with version numbers (e.g., 'Python 3.8, PyTorch 1.9').
Experiment Setup No The paper describes the model architecture (Res Net with specific layer/filter details) and attack parameters (shape, position, color, epsilon), but does not provide specific training hyperparameters such as learning rate, batch size, or number of epochs.