Parallel and Streaming Algorithms for K-Core Decomposition

Authors: Hossein Esfandiari, Silvio Lattanzi, Vahab Mirrokni

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

Reproducibility Variable Result LLM Response
Research Type Experimental Finally, we confirm the effectiveness of our sketching technique empirically on a number of publicly available graphs.
Researcher Affiliation Industry 1Google Research. Correspondence to: Hossein Esfandiari <esfandiari@googol.com>.
Pseudocode Yes Algorithm 1 Exclusive Coreness Labeling(H, Λ)
Open Source Code No The paper describes the implementation of the algorithms and experimental evaluation, but it does not provide any statement or link indicating that the source code for the methodology is openly available.
Open Datasets Yes We apply our sketch to eight real-world graphs available in the SNAP, Stanford Large Network Dataset Library (Leskovec & Sosiˇc, 2016): Enron (Klimt & Yang, 2004), Epinions (Richardson et al., 2003), Slashdot (Leskovec et al., 2009), Twitter (Mc Auley & Leskovec, 2012), Amazon (Yang & Leskovec, 2015), Youtube (Yang & Leskovec, 2015), Live Journal (Yang & Leskovec, 2015), Orkut (Yang & Leskovec, 2015)
Dataset Splits No The paper mentions using real-world graphs for experiments but does not specify any train/validation/test splits or reference standard splits from the datasets.
Hardware Specification No The paper discusses distributed and parallel algorithms, but it does not provide any specific details about the hardware (e.g., CPU/GPU models, memory, cluster specifications) used for running the experiments.
Software Dependencies No The paper describes algorithms and their implementation but does not specify any software dependencies with version numbers (e.g., programming languages, libraries, or frameworks).
Experiment Setup Yes More specifically, we change line 14 to if lj(i) T pj = 1 where T is a parameter of our implementation. Furthermore, we also modify line 22 to pj+1 M pj, where M is modifiable multiplicative factor (that in Algorithm 2 is fixed to 2). We also slightly modify our Map Reduce algorithm to remove iteratively in parallel all nodes with degree smaller than 3 before sending the remaining graph to a single machine.