Fast Certified Robust Training with Short Warmup

Authors: Zhouxing Shi, Yihan Wang, Huan Zhang, Jinfeng Yi, Cho-Jui Hsieh

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

Reproducibility Variable Result LLM Response
Research Type Experimental We are able to obtain 65.03% verified error on CIFAR-10 ( = 8 255) and 82.36% verified error on Tiny Image Net ( = 1 255) using very short training schedules (160 and 80 total epochs, respectively), outperforming literature SOTA trained with hundreds or thousands epochs under the same network architecture. The code is available at https: //github.com/shizhouxing/Fast-Certified-Robust-Training.
Researcher Affiliation Collaboration 1University of California, Los Angeles 2Carnegie Mellon University 3JD AI Research
Pseudocode No The paper does not contain any pseudocode or algorithm blocks.
Open Source Code Yes The code is available at https: //github.com/shizhouxing/Fast-Certified-Robust-Training.
Open Datasets Yes We adopt three datasets, MNIST (Le Cun et al., 2010), CIFAR-10 (Krizhevsky et al., 2009) and Tiny Image Net (Le & Yang, 2015).
Dataset Splits No The paper mentions training epochs and datasets, but does not explicitly specify training/validation/test dataset splits or validation set sizes within the main text or appendices.
Hardware Specification Yes We compare the training cost using a single NVIDIA RTX 2080 Ti GPU.
Software Dependencies No The paper mentions software like 'Py Torch' and 'Adam optimizer', but it does not provide specific version numbers for these or other key software components.
Experiment Setup Yes For all models, we train for 160 epochs on CIFAR-10, 70 epochs on MNIST, and 80 epochs on Tiny Image Net... We use Adam optimizer (Kingma & Ba, 2014) with β1 = 0.9, β2 = 0.999. The initial learning rate is 5e-4 and decays following cosine annealing schedule...