Towards Understanding the Dynamics of the First-Order Adversaries
Authors: Zhun Deng, Hangfeng He, Jiaoyang Huang, Weijie Su
ICML 2020 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | Our analysis demonstrates that, in the initial phase of adversarial training, the scale of the inputs matters in the sense that a smaller input scale leads to faster convergence of adversarial training and a more regular landscape. Finally, we show that these theoretical findings are in excellent agreement with a series of experiments. |
| Researcher Affiliation | Academia | 1Harvard University 2University of Pennsylvania 3Institute of Advanced Study. |
| Pseudocode | No | The paper provides mathematical equations for the projected gradient ascent update rule (e.g., 'δt+1 = PB(0,ε) h δt + η L(δt)'), but it does not present these or any other procedural steps in a formally structured pseudocode or algorithm block. |
| Open Source Code | No | The paper does not contain any explicit statements about making its source code publicly available, nor does it provide links to a code repository for the methodology described. |
| Open Datasets | Yes | The experiments are based on a real-world dataset MNIST and a practical multi-layer CNN. |
| Dataset Splits | No | The paper mentions using the MNIST dataset for experiments, but it does not provide specific details about training, validation, or test dataset splits (e.g., percentages, sample counts, or references to predefined splits). |
| Hardware Specification | No | The paper does not provide any specific details about the hardware (e.g., GPU models, CPU types, memory) used to conduct the experiments. |
| Software Dependencies | No | The paper does not specify any software dependencies with version numbers (e.g., Python version, library names with versions) that would be needed to replicate the experiments. |
| Experiment Setup | No | The paper states that parameters are taken 'according to Xavier initialization' and discusses the impact of learning rate and input scales, but it does not provide concrete numerical values for hyperparameters (e.g., specific learning rates, batch sizes, number of epochs) or system-level training settings in the main text. It defers some details to supplementary materials. |