A Causal View on Robustness of Neural Networks

Authors: Cheng Zhang, Kun Zhang, Yingzhen Li

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

Reproducibility Variable Result LLM Response
Research Type Experimental Compared to DNNs, experiments on both MNIST and a measurement-based dataset show that our model is significantly more robustness to unseen manipulations.
Researcher Affiliation Collaboration Cheng Zhang Microsoft Research Cheng.Zhang@microsoft.com Kun Zhang Carnegie Mellon University kunz1@cmu.edu Yingzhen Li Microsoft Research Yingzhen.Li@microsoft.com
Pseudocode No The paper does not contain structured pseudocode or algorithm blocks (e.g., sections explicitly labeled "Algorithm" or "Pseudocode").
Open Source Code No The paper does not provide an explicit statement about the release of source code or a link to a code repository.
Open Datasets Yes We evaluate the robustness of deep CAMA for image classification using both MNIST and a binary classification task derived from CIFAR-10
Dataset Splits No The paper discusses training and testing, but does not provide specific details on validation dataset splits (e.g., percentages, sample counts, or a clear methodology for validation splits).
Hardware Specification No The paper mentions "Nathan Jones for his support with computing infrastructure" but does not specify any particular hardware details such as GPU models, CPU types, or memory amounts used for the experiments.
Software Dependencies No The paper does not provide specific version numbers for any software dependencies or libraries used (e.g., Python, PyTorch, TensorFlow, etc.).
Experiment Setup No The paper describes general training scenarios (clean data, augmented data) and fine-tuning, but does not provide specific experimental setup details such as concrete hyperparameter values (e.g., learning rate, batch size, number of epochs) or optimizer settings.