MACER: Attack-free and Scalable Robust Training via Maximizing Certified Radius

Authors: Runtian Zhai, Chen Dan, Di He, Huan Zhang, Boqing Gong, Pradeep Ravikumar, Cho-Jui Hsieh, Liwei Wang

ICLR 2020 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental In our experiments, we show that our method can be applied to modern deep neural networks on a wide range of datasets, including Cifar-10, Image Net, MNIST, and SVHN.
Researcher Affiliation Collaboration 1Peking University 2CMU 3UCLA 4Google
Pseudocode Yes Algorithm 1 MACER: robust training via MAximizing CErtified Radius
Open Source Code Yes Our code is available at https://github.com/Runtian Z/macer.
Open Datasets Yes In our experiments, we show that our method can be applied to modern deep neural networks on a wide range of datasets, including Cifar-10, Image Net, MNIST, and SVHN.
Dataset Splits No The paper mentions a training set and a test set, but does not explicitly provide details for a validation set split (e.g., percentages or sample counts).
Hardware Specification Yes For Cifar-10 we use one NVIDIA P100 GPU and for Image Net we use four NVIDIA P100 GPUs.
Software Dependencies No The paper does not specify software dependencies with version numbers (e.g., programming languages, libraries, or frameworks with their specific versions).
Experiment Setup Yes For Cifar-10, MNIST and SVHN, we train the models for 440 epochs using our proposed algorithm. The learning rate is initialized to be 0.01, and is decayed by 0.1 at the 200th/400th epoch. For all the models, we use k = 16, γ = 8.0 and β = 16.0.