Posterior Collapse and Latent Variable Non-identifiability

Authors: Yixin Wang, David Blei, John P. Cunningham

NeurIPS 2021 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental Across synthetic and real datasets, latent-identifiable variational autoencoders outperform existing methods in mitigating posterior collapse and providing meaningful representations of the data. ... 4 Empirical studies
Researcher Affiliation Academia Yixin Wang University of Michigan yixinw@umich.edu David M. Blei Columbia University david.blei@columbia.edu John P. Cunningham Columbia University jpc2181@columbia.edu
Pseudocode No The paper describes the models and methods using mathematical equations and textual explanations, but it does not include any clearly labeled pseudocode or algorithm blocks.
Open Source Code No The paper does not provide a specific link to a source-code repository or an explicit statement confirming the release of source code for the methodology described.
Open Datasets Yes We first study LIDGMVAE on four subsampled image datasets drawn from pinwheel [22], MNIST [31], Fashion MNIST [57], and Omniglot [30]. ... We apply LIDSVAE to three subsampled text datasets drawn from a synthetic text dataset, the Yahoo dataset, and the Yelp dataset [60].
Dataset Splits No The paper mentions evaluating on a 'held-out test set' but does not provide specific details on the training, validation, and test splits (e.g., percentages, sample counts, or predefined split references).
Hardware Specification No The paper discusses the computational cost and complexity of the model but does not provide any specific details regarding the hardware (e.g., GPU/CPU models, memory specifications) used for running the experiments.
Software Dependencies No The paper mentions software components and techniques such as 'Real NVPs', 'LSTMs', and 'Hamiltonian Monte Carlo', but does not provide specific version numbers for any of these software dependencies.
Experiment Setup No The paper mentions using a '9-layer generative model' and discusses various aspects of the models, but it does not explicitly provide specific hyperparameter values (e.g., learning rate, batch size, number of epochs) or detailed training configurations in the main text.