Importance Weighting and Variational Inference

Authors: Justin Domke, Daniel R. Sheldon

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

Reproducibility Variable Result LLM Response
Research Type Experimental All the following experiments compare E-IWVI using student T distributions to IWVI using Gaussians.
Researcher Affiliation Academia 1 College of Information and Computer Sciences, University of Massachusetts Amherst 2 Department of Computer Science, Mount Holyoke College
Pseudocode Yes Algorithm 1 A generative process for q M(z1:M)
Open Source Code No The paper does not provide an unambiguous statement or a link to open-source code for the methodology described.
Open Datasets Yes From top: Madelon (d = 500) Sonar (d = 60), Mushrooms (d = 112).
Dataset Splits No The paper does not provide specific dataset split information (exact percentages, sample counts, or detailed splitting methodology) needed to reproduce the data partitioning.
Hardware Specification No The paper does not provide specific hardware details (exact GPU/CPU models, processor types, or memory amounts) used for running its experiments.
Software Dependencies Yes Stan Development Team. Modeling language user s guide and reference manual, version 2.17.0
Experiment Setup Yes On these, we used a fixed set of 10, 000 M random inputs to T and optimized using batch L-BFGS, avoiding heuristic tuning of a learning rate sequence. Finally, we considered a (non-conjugate) logistic regression model with a Cauchy prior with a scale of 10, using stochastic gradient descent with various step sizes.