Bipartite Stochastic Block Models with Tiny Clusters

Authors: Stefan Neumann

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

Reproducibility Variable Result LLM Response
Research Type Experimental We evaluate the algorithm on synthetic and on real-world data; the experiments show that the algorithm can find extremely small clusters even in presence of high destructive noise.
Researcher Affiliation Academia Stefan Neumann University of Vienna Faculty of Computer Science Vienna, Austria stefan.neumann@univie.ac.at
Pseudocode Yes Algorithm 1 The pcv algorithm Input: G a bipartite m n graph, k, p, q
Open Source Code Yes We implemented Algorithm 1 in Python. To compute the truncated SVD we used scikit-learn [23]. The source code is available in the supplementary material.
Open Datasets Yes The source code and the synthetic data are provided in the supplementary materials. and The Book Crossing dataset1 originates from Ziegler et al. [30].
Dataset Splits No No explicit statement of dataset splits (e.g., train/validation percentages or counts) was found. The paper mentions generating synthetic data and evaluating on random graphs, but not how a single dataset was partitioned for training, validation, and testing.
Hardware Specification Yes The experiments were done on a Mac Book Air with a 1.6 GHz Intel Core i5 and 8 GB RAM.
Software Dependencies No The paper mentions 'Python' and 'scikit-learn [23]' but does not provide specific version numbers for these software components.
Experiment Setup Yes When not mentioned otherwise, the parameters were set to n = 1000, k = 8, ℓ= 70, and m = ℓ k (i.e., 1000 vertices on the right, 8 ground-truth clusters on both sides and left-side clusters of size 70). The size of the right-side clusters was set to r = 8. The parameters p and q were set depending on the dataset.