Sublinear-Time Clustering Oracle for Signed Graphs

Authors: Stefan Neumann, Pan Peng

ICML 2022 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental We evaluate our algorithm for constructing such an oracle and answering membership queries on both synthetic and real-world datasets, validating its performance in practice. and We evaluate our algorithms on synthetic and on real-world datasets (Sec. 4) and show that our oracles are practical.
Researcher Affiliation Academia 1KTH Royal Institute of Technology, Stockholm, Sweden 2School of Computer Science and Technology, University of Science and Technology of China, Hefei, China.
Pseudocode Yes Algorithm 1 Estimating the dot product pt u D 1/2, pt v D 1/2, Algorithm 2 Preprocessing: Constructing a clustering oracle, Algorithm 3 Answering the community membership of a vertex v
Open Source Code Yes Our source code is available on github.2 with footnote 2https://github.com/Stefan Research/ signed-oracle
Open Datasets Yes We created three dataset. Wiki L contains all politicans and all articles linked on their pages; we included all edges that contain at least one politician... We make them available on github.2 with footnote 2https://github.com/Stefan Research/ signed-oracle.
Dataset Splits No The paper describes how seed nodes are selected (e.g., 'the seeded oracles obtained 10 (5) seeds for each Ui (Vi); the unseeded oracles sampled 5k vertices') but does not specify explicit training, validation, or test dataset splits of the overall datasets used for evaluation.
Hardware Specification Yes We experimentally evaluated our algorithms on a Mac Book Pro with 16 GB RAM and a 2 GHz Quad-Core Intel Core i5.
Software Dependencies No Our algorithms were implemented in C++11. (While a specific language version is given, no other software dependencies or libraries with version numbers are mentioned to ensure reproducibility of the environment.)
Experiment Setup Yes Unless stated otherwise, our oracles used 1000 random walks of length 20. and We ran the oracles with t = 2 random walk steps and R = 400 random walks, unless stated otherwise.