Code associated with Concentration of Multilinear Functions of the Ising Model with Applications to Network Data, by Constantinos Daskalakis, Nishanth Dikkala, and Gautam Kamath. Paper appeared at Advances in Neural Information Processing Systems 30 (NIPS 2017).

Implementation is in MATLAB. To run processMusic.m, one requires the HetRec 2011 dataset. This involves extracting hetrec2011-lastfm-2k.zip to the directory multilinear-ising/data.

Figures from the paper can be generated by running the following files:

• Figure 1: visualize.m.
• Figure 2: syntheticExperiment.m. Loads a precomputed set of MCMC statistics, which can be regenerated by running syntheticPrecomputeStats.m
• Figure 3 and 4: processMusic.m. Requires uncommenting the relevant lines in the script.

Bibliographic Information

If you use our code or paper, we ask that you please cite:

@inproceedings{DaskalakisDK17,
  author        = {Daskalakis, Constantinos and Dikkala, Nishanth and Kamath, Gautam},
  title         = {Concentration of Multilinear Functions of the {I}sing Model with Applications to Network Data},
  booktitle     = {Advances in Neural Information Processing Systems 30},
  series        = {NIPS '17},
  year          = {2017},
  pages         = {12--23},
  publisher     = {Curran Associates, Inc.}
}