On Unifying Deep Generative Models

Authors: Zhiting Hu, Zichao Yang, Ruslan Salakhutdinov, Eric P. Xing

ICLR 2018 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental We conduct preliminary experiments to demonstrate the generality and effectiveness of the importance weighting (IW) and adversarial activating (AA) techniques. In this paper we do not aim at achieving state-of-the-art performance, but leave it for future work. In particular, we show the IW and AA extensions improve the standard GANs and VAEs, as well as several of their variants, respectively. We present the results here, and provide details of experimental setups in the supplements.
Researcher Affiliation Collaboration Zhiting Hu1,2 Zichao Yang1 Ruslan Salakhutdinov1 Eric P. Xing1,2 Carnegie Mellon University1, Petuum Inc.2
Pseudocode No No pseudocode or clearly labeled algorithm block was found in the paper.
Open Source Code No The paper does not include an unambiguous statement or a direct link indicating the release of open-source code for the described methodology.
Open Datasets Yes We use MNIST, SVHN, and CIFAR10 for evaluation.
Dataset Splits No The paper mentions 'test set' and varying 'Train Data Size' (1%, 10%, 100%) for MNIST, and 'cross-validation' for hyperparameter tuning, but it does not provide explicit percentages or sample counts for the training, validation, and test splits needed for reproduction.
Hardware Specification No The paper does not provide any specific hardware details (e.g., GPU/CPU models, memory amounts) used for running its experiments.
Software Dependencies No The paper mentions 'tensorflow library' but does not provide specific version numbers for software dependencies needed to replicate the experiment.
Experiment Setup Yes The base GAN model is implemented with the DCGAN architecture and hyperparameter setting (Radford et al., 2015). Hyperparameters are not tuned for the IW extensions. We select the best temperature from {1, 1.5, 3, 5} through cross-validation.