Copula variational inference

Authors: Dustin Tran, David Blei, Edo M. Airoldi

NeurIPS 2015 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental In experiments, we demonstrate copula vi on the standard example of Gaussian mixture models. We study copula vi with two models: Gaussian mixtures and the latent space model [7].
Researcher Affiliation Academia Dustin Tran Harvard University David M. Blei Columbia University Edoardo M. Airoldi Harvard University
Pseudocode Yes Algorithm 1: Copula variational inference (copula vi)
Open Source Code No The paper states 'The rest of the algorithm s calculations, such as sampling and evaluating gradients, can be placed in a library.' and 'Thus Copula vi can be implemented in a library and applied without requiring any manual derivations from the user.' but does not provide concrete access to its own source code.
Open Datasets Yes We analyze the MNIST data set of handwritten digits, using 12,665 training examples and 2,115 test examples of 0 s and 1 s.
Dataset Splits No The paper mentions '12,665 training examples and 2,115 test examples' for the MNIST dataset but does not provide explicit details about validation splits or general data partitioning strategy for reproducibility across all experiments.
Hardware Specification No No specific hardware details (exact GPU/CPU models, processor types with speeds, memory amounts, or detailed computer specifications) used for running experiments were found.
Software Dependencies No The paper mentions using 'ADAM [12]' for step-size and 'automatic differentiation tools [22]' but does not provide specific version numbers for any software dependencies.
Experiment Setup Yes We set m = 1024 and follow asynchronous updates [16]. We set the step-size using ADAM [12].