Beyond the Birkhoff Polytope: Convex Relaxations for Vector Permutation Problems

Authors: Cong Han Lim, Stephen Wright

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

Reproducibility Variable Result LLM Response
Research Type Experimental We provide experimental results showing the effectiveness of our approach.
Researcher Affiliation Academia Cong Han Lim Department of Computer Sciences University of Wisconsin Madison Madison, WI 53706 conghan@cs.wisc.edu Stephen J. Wright Department of Computer Sciences University of Wisconsin Madison Madison, WI 53706 swright@cs.wisc.edu
Pseudocode No The paper describes algorithms and formulations but does not include any explicit pseudocode blocks or labeled algorithm sections.
Open Source Code No The paper does not provide concrete access to its own source code for the methodology described. It mentions using "Experimental MATLAB code provided to us by the authors of [2]" but this refers to third-party code.
Open Datasets No The authors generated their own data for experiments: "We fixed b = 0.999 and σ = 0.5 and generate 50 chains to form a sample covariance matrix." This is not a publicly available dataset with concrete access information.
Dataset Splits No The paper describes generating different sets of samples for independent runs but does not specify traditional training, validation, or test dataset splits. "For each n, we perform 10 independent runs, each based on a different set of samples of the Markov chain."
Hardware Specification Yes The experiments were run on an Intel Xeon X5650 (24 cores @ 2.66Ghz) server with 128GB of RAM in MATLAB 7.13, CVX 2.0 ([14],[15]), and Gurobi 5.5 [16].
Software Dependencies Yes The experiments were run on an Intel Xeon X5650 (24 cores @ 2.66Ghz) server with 128GB of RAM in MATLAB 7.13, CVX 2.0 ([14],[15]), and Gurobi 5.5 [16].
Experiment Setup Yes For both sets of experiments, we fixed b = 0.999 and σ = 0.5 and generate 50 chains to form a sample covariance matrix. ... We used a regularization parameter of µ = 0.9λ2(LA) on all formulations. ... For RQPS, with a cap of 10 iterations within each projection step, ... For a limit of 100 iterations...