Finding Dense Subgraphs via Low-Rank Bilinear Optimization

Authors: Dimitris Papailiopoulos, Ioannis Mitliagkas, Alexandros Dimakis, Constantine Caramanis

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

Reproducibility Variable Result LLM Response
Research Type Experimental We demonstrate this by implementing it in Map Reduce and executing numerous experiments on massive real-world graphs that have up to billions of edges. We empirically show that our algorithm can find subgraphs of significantly higher density compared to the previous state of the art.
Researcher Affiliation Academia Dimitris S. Papailiopoulos DIMITRIS@UTEXAS.EDU Ioannis Mitliagkas IOANNIS@UTEXAS.EDU Alexandros G. Dimakis DIMAKIS@AUSTIN.UTEXAS.EDU Constantine Caramanis CONSTANTINE@UTEXAS.EDU The University of Texas at Austin
Pseudocode Yes The exact steps of our algorithm are given in the pseudocode tables referred to as Algorithms 1-3.
Open Source Code No The paper mentions using open-source tools like 'MRJob' and 'Amazon Web Services Elastic Map Reduce framework' with URLs, but it does not state that the code for the DkS algorithm developed in this paper is open-source or provide a link to its own implementation.
Open Datasets No The paper mentions using 'massive real-world graphs' and specific datasets such as 'com-Live Journal', 'com-DBLP', 'web Notre Dame', and a 'Facebook graph' in the experimental section, but it does not provide concrete access information (e.g., specific links, DOIs, repositories, or formal citations with authors/year) for these datasets.
Dataset Splits No The paper does not provide specific percentages or counts for train/validation/test dataset splits. It discusses experiments on various graphs but not the data partitioning methodology.
Hardware Specification Yes For our biggest experiments we used a 100-machine strong cluster, consisting of m1.xlarge AWS instances (a total of 800 cores). ... all performed on a standard Macbook Pro laptop using Matlab.
Software Dependencies No The paper mentions implementing functions in 'Python' and using the 'MRJob class (mrj)', as well as 'Matlab'. However, it does not provide specific version numbers for any of these software components, which is required for reproducibility.
Experiment Setup No The paper does not provide specific hyperparameter values (e.g., learning rate, batch size, number of epochs) or detailed system-level training settings. It mentions 'different rank choices' and 'few (typically 4) iterations' for power iteration, but these are not comprehensive experimental setup details.