A Deterministic Partition Function Approximation for Exponential Random Graph Models

Authors: Wen Pu, Jaesik Choi, Yunseong Hwang, Eyal Amir

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

Reproducibility Variable Result LLM Response
Research Type Experimental Our results show that the new method performs well experimentally comparing to existing sampling methods [Gelman and Meng, 1998; Handcock et al., 2003] on synthetic data and real-world social networks.
Researcher Affiliation Collaboration Wen Pu LinkedIn Corporation Mountain View, CA, USA wpu@linkedin.com Jaesik Choi Ulsan National Institute of Science and Technology, Ulsan, Korea jaesik@unist.ac.kr Yunseong Hwang Ulsan National Institute of Science and Technology, Ulsan, Korea yunseong@unist.ac.kr Eyal Amir University of Illinois at Urbana-Champaign Urbana, IL, USA eyal@illinois.edu
Pseudocode Yes Algorithm 1 Our new ECS Approximation to the log partition function ln Z(θ) Input: model parameter θ and number of nodes n Output: estimation of ln Z(θ) Initialize ECS for u 0 to n(n 1)/2 do ECS max{ γ(θ, u), ECS} end for
Open Source Code No No statement regarding the release of open-source code for the described methodology is provided, nor is a link to a code repository.
Open Datasets Yes To study the stability and quality of ECS-MLE results, we fit two different models with four networks with more than 40 nodes: one kapferer2 from statnet; and the other three networks, prison, dolphins, and sanjuansur, from CMU CASOS.
Dataset Splits No No specific training/test/validation splits are provided. The paper describes generating synthetic data and simulating graphs from models for comparison, but does not detail dataset partitioning for reproduction.
Hardware Specification No No specific hardware details (e.g., CPU, GPU models, memory, or cluster specifications) used for running the experiments are mentioned in the paper.
Software Dependencies No The paper mentions software like 'statnet' and algorithms such as 'Bridge Sampling' and 'MCMC-MLE', but does not provide specific version numbers for any software dependencies required to replicate the experiments.
Experiment Setup Yes For ECS-MLE, we performed grid search in a slightly enlarged parameter space with finer granularity. For triad models, our ECS algorithm searched the grid of two parameters of (edge, 2-star, triangle) ranging the initial value plus offsets in [ -5, 5] x [ -5, 5] x [ -5, 5], with granularity of 0.2. For alternating k-stars, the ECS searched the grid of three parameters of (edge, altkstar) ranging, initial value plus offsets in [ -15, 15] x [ -15, 15], with the same granularity.