On the Global Convergence of (Fast) Incremental Expectation Maximization Methods

Authors: Belhal Karimi, Hoi-To Wai, Eric Moulines, Marc Lavielle

NeurIPS 2019 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental Numerical applications are presented in this article to illustrate our findings. and 4 Numerical Examples
Researcher Affiliation Academia Belhal Karimi CMAP, Ecole Polytechnique Palaiseau, France belhal.karimi@polytechnique.edu, Hoi-To Wai The Chinese University of Hong Kong Shatin, Hong Kong htwai@se.cuhk.edu.hk, Eric Moulines CMAP, Ecole Polytechnique Palaiseau, France eric.moulines@polytechnique.edu, Marc Lavielle INRIA Saclay Palaiseau, France marc.lavielle@inria.fr
Pseudocode Yes Algorithm 1 Stochastic EM methods.
Open Source Code No The paper does not provide concrete access to source code for the methodology described, such as a specific repository link or an explicit code release statement.
Open Datasets Yes We compare the stochastic EM methods on two FAO (UN Food and Agriculture Organization) datasets [Medelyan, 2009].
Dataset Splits No The paper mentions using synthetic data and FAO datasets but does not provide specific dataset split information (exact percentages, sample counts, citations to predefined splits, 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 with speeds, memory amounts, or detailed computer specifications) used for running its experiments.
Software Dependencies No The paper does not provide specific ancillary software details (e.g., library or solver names with version numbers) needed to replicate the experiment.
Experiment Setup Yes We set the stepsize of the s EM as γk = 3/(k + 10), and the stepsizes of the s EM-VR and the fiEM to a constant stepsize proportional to 1/n2/3 and equal to γ = 0.003. and The number of topics is set to K = 10 and the stepsizes for the fiEM and s EM-VR are set to γ = 1/n2/3 while the stepsize for the s EM is set to γk = 1/(k + 10).