LaVAN: Localized and Visible Adversarial Noise

Authors: Danny Karmon, Daniel Zoran, Yoav Goldberg

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

Reproducibility Variable Result LLM Response
Research Type Experimental We experiment with two setups: working in the Network Domain and in the Image Domain. 2.2. Experiments and Results We use the Py Torch provided pre-trained Inception V3 network (Szegedy et al., 2016) trained on Image Net.
Researcher Affiliation Collaboration 1Microsoft, Herzliya, Israel. 2Deep Mind, London, UK. 3Department of Computer Science, Bar-Ilan University,Ramat Gan, Israel.
Pseudocode Yes Algorithm 1 Localized Noising Process
Open Source Code No The paper does not provide any specific repository links or explicit statements about releasing the source code for the described methodology.
Open Datasets Yes We use the Py Torch provided pre-trained Inception V3 network (Szegedy et al., 2016) trained on Image Net.
Dataset Splits No The paper mentions a “training set of 100 images” and refers to a “separate test set, consisting of 100 images”, but does not provide specific details for validation dataset splits (e.g., percentages, counts, or explicit use of a validation set for hyperparameter tuning) needed for reproduction.
Hardware Specification No The paper does not provide specific hardware details such as GPU/CPU models, processor types, or memory specifications used for running its experiments.
Software Dependencies No The paper mentions “Py Torch” and “Inception V3 network” but does not specify exact version numbers for these or other software dependencies.
Experiment Setup Yes We use images of size 299x299 and noise a square patch of size 42x42, roughly 2% of the image pixels. For each image, we train the noise until we reach the desired confidence, or up to 10,000 iterations. Target class probability 0.9