Dropout Training, Data-dependent Regularization, and Generalization Bounds
Authors: Wenlong Mou, Yuchen Zhou, Jun Gao, Liwei Wang
ICML 2018 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | In this section, we conduct experiments to verify the effectiveness of Algorithm 1, the truthful dropout, demonstrating the comparable generalization improvement in terms of classification error when comparing with standard Dropout. We use MNIST (Le Cun et al., 1998) and CIFAR10 (Krizhevsky & Hinton, 2009) datasets to test our algorithm, with both convolutional and fully-connected neural networks. The details about experimental setup are postponed to the Appendix. The classification error on test set is shown in the Table 1. We also plot the curve for classification error during optimization in the Figure 1. |
| Researcher Affiliation | Academia | 1Department of EECS, University of California, Berkeley 2Department of Statistics, University of Wisconsin, Madison 3Key Laboratory of Machine Perception, MOE, School of EECS, Peking University 4Center for Data Science, Peking University, Beijing Institute of Big Data Research. |
| Pseudocode | Yes | Algorithm 1 Truthful dropout feed-forwarding |
| Open Source Code | No | The paper does not contain any explicit statement or link indicating that the source code for the described methodology is publicly available. |
| Open Datasets | Yes | We use MNIST (Le Cun et al., 1998) and CIFAR10 (Krizhevsky & Hinton, 2009) datasets to test our algorithm... |
| Dataset Splits | No | The paper mentions training and test sets but does not explicitly state details about validation dataset splits (e.g., percentages, sample counts, or specific strategies for a validation set). |
| Hardware Specification | No | The paper does not provide specific details about the hardware (e.g., GPU models, CPU types, memory) used to run the experiments. |
| Software Dependencies | No | The paper does not list specific software dependencies with version numbers (e.g., Python 3.x, TensorFlow x.x, PyTorch x.x). |
| Experiment Setup | No | The paper states: 'The details about experimental setup are postponed to the Appendix.' Therefore, specific details are not provided in the main text. |