Multi-Scale Spectral Decomposition of Massive Graphs

Authors: Si Si, Donghyuk Shin, Inderjit S Dhillon, Beresford N Parlett

NeurIPS 2014 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental Experimental results are shown in Section 4 followed by conclusions in Section 5.
Researcher Affiliation Academia Si Si Department of Computer Science University of Texas at Austin ssi@cs.utexas.edu Donghyuk Shin Department of Computer Science University of Texas at Austin dshin@cs.utexas.edu Inderjit S. Dhillon Department of Computer Science University of Texas at Austin inderjit@cs.utexas.edu Beresford N. Parlett Department of Mathematics University of California, Berkeley parlett@math.berkeley.edu
Pseudocode Yes Algorithm 1: MSEIGS with single level
Open Source Code No The paper mentions using open-source tools like Metis and Graclus, but it does not explicitly state that the source code for the proposed MSEIGS method itself is openly available or provide a link to a repository.
Open Datasets Yes Summary of the datasets is given in Table 1, where the largest graph contains more than 3.6 billion edges. We use the average of the cosine of principal angles cos( ( Uk, Uk)) as the evaluation metric... (a) Cond Mat (b) Amazon (c) Road CA (d) Live Journal (e) Friendster Sub (f) SDWeb. In Table 2, we compare MSEIGS and MSEIGS-Early with other methods for label propagation on two public datasets: Aloi and Delicious... The datasets are referenced with citations such as [17, 28] and [16].
Dataset Splits Yes We evaluated the recommendation performance on three publicly available datasets shown in Table 6 (see Appendix 6.7 for more details). We report recall-at-N with N = 20 averaged over 5-fold cross-validation, which is a widely used evaluation metric for top-N recommendation tasks [2].
Hardware Specification No The paper states, 'on a single-core machine' and 'Using 16 cores' but does not specify any particular hardware components such as CPU or GPU models, processor types, or memory specifications.
Software Dependencies No The paper mentions software tools like 'Matlab s eigs function (EIGS) [14]', 'PROPACK [12]', 'randomized SVD (RSVD) [7]', 'block Lanczos (Blk Lan) [21]', and 'Intel MKL', as well as clustering software such as 'Graclus [5], Metis [11], Nerstrand [13] and GEM [27]', but it does not provide specific version numbers for these components.
Experiment Setup No The paper states, 'The experimental settings can be found in Appendix 6.5.' indicating that specific setup details like hyperparameters are not present in the main body of the paper.