PacGAN: The power of two samples in generative adversarial networks
Authors: Zinan Lin, Ashish Khetan, Giulia Fanti, Sewoong Oh
NeurIPS 2018 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | Numerical experiments on benchmark datasets suggest that packing provides significant improvements. We demonstrate on benchmark datasets that Pac GAN significantly improves upon competing approaches in mitigating mode collapse (Section 4), notably minibatch discrimination [24]. |
| Researcher Affiliation | Academia | Zinan Lin ECE Department Carnegie Mellon University zinanl@andrew.cmu.edu Ashish Khetan IESE Department University of Illinois at Urbana-Champaign ashish.khetan09@gmail.com Giulia Fanti ECE Department Carnegie Mellon University gfanti@andrew.cmu.edu Sewoong Oh IESE Department University of Illinois at Urbana-Champaign swoh@illinois.edu |
| Pseudocode | No | The paper does not contain any clearly labeled pseudocode or algorithm blocks. It describes processes in text and uses mathematical equations, but no structured pseudocode. |
| Open Source Code | Yes | All code to reproduce our experiments can be found at https://github.com/fjxmlzn/PacGAN. |
| Open Datasets | Yes | Datasets. We use synthetic and real datasets. The 2D-ring [25]... The 2D-grid [25]... The MNIST dataset [16] consists of 70K images of handwritten digits, each 28 28 pixels... Finally, we include experiments on the Celeb A dataset, which is a collection of 200K facial images of celebrities [19]. |
| Dataset Splits | No | The paper mentions using datasets for training but does not explicitly provide specific training, validation, or test dataset splits (e.g., percentages, sample counts, or references to predefined splits) for its own experiments in the main text. |
| Hardware Specification | No | The paper mentions using "the Extreme Science and Engineering Discovery Environment (XSEDE)", the "Bridges system", and "AWS cloud computing resources". However, it does not provide specific hardware details such as GPU models (e.g., NVIDIA V100), CPU models, or memory specifications. |
| Software Dependencies | No | The paper mentions "DCGAN-tensorflow" in a footnote related to an implementation, implying TensorFlow was used. However, it does not provide specific version numbers for any software, libraries, or dependencies used in their experiments. |
| Experiment Setup | No | The paper states: "The experimental details are included in Appendix A.2." and "details in Appendix B.1". It defers specific hyperparameter values and training configurations to the appendices rather than providing them in the main text. |