Maximum Likelihood Learning With Arbitrary Treewidth via Fast-Mixing Parameter Sets

Authors: Justin Domke

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

Reproducibility Variable Result LLM Response
Research Type Experimental Figure 2: Ising Model Example. Left: The difference of the current test log-likelihood from the optimal log-likelihood on 5 random runs.
Researcher Affiliation Collaboration Justin Domke NICTA, Australian National University
Pseudocode Yes Figure 1: Left: Algorithm 1, approximate gradient descent with gradients approximated via MCMC, analyzed in this paper.
Open Source Code No The paper does not provide concrete access to source code for the methodology described.
Open Datasets No The paper describes using '5 random vectors as training data' for an Ising model example, but no concrete access information (link, DOI, repository, formal citation) for a publicly available dataset is provided.
Dataset Splits No The paper does not provide specific dataset split information (exact percentages, sample counts, citations to predefined splits, or detailed splitting methodology).
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 No The paper does not provide specific ancillary software details with version numbers.
Experiment Setup Yes For a fast-mixing set, constrain |θij| .2 for all pairs. Since the maximum degree is 4, τ(ϵ) N log(N/ϵ) / (1 - 4 tanh(.2)). Fix λ = 1, ϵθ = .2 and δ = 0.1. Though the theory above suggests the Lipschitz constant L = 4R2 + λ = 97, a lower value of L = 10 is used, which converged faster in practice (with exact or approximate gradients).