Adaptive Clustering through Semidefinite Programming
Authors: Martin Royer
NeurIPS 2017 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | We illustrate the method s performances in comparison to other classical clustering algorithms with numerical experiments on simulated high-dimensional data. |
| Researcher Affiliation | Academia | Martin Royer Laboratoire de Mathématiques d Orsay, Univ. Paris-Sud, CNRS, Université Paris-Saclay, 91405 Orsay, France martin.royer@math.u-psud.fr |
| Pseudocode | No | The paper describes various algorithms and estimators but does not include any structured pseudocode or algorithm blocks. |
| Open Source Code | Yes | The PYTHON3 implementation of the method used is found in open access here: martinroyer/pecok [18] |
| Open Datasets | No | Our sample of n = 100 points are drawn from K = 5 identically-sized, perfectly discriminated non-isovolumic clusters of Gaussians that is we have k [K], a Gk, Ea N(0, Σk) such that |G1| = ... = |GK| = 20. The paper uses simulated data, and thus does not refer to a publicly available dataset. |
| Dataset Splits | No | The paper uses simulated data and describes two experimental scenarios (S1 and S2) with "a hundred simulations" each. It does not mention explicit training, validation, or test splits, nor does it describe a cross-validation setup. |
| Hardware Specification | No | The paper mentions computation times but does not provide specific hardware details such as GPU or CPU models, or memory specifications used for the experiments. |
| Software Dependencies | No | The paper mentions "PYTHON3 implementation" but does not provide specific version numbers for Python or any other key software dependencies or libraries used in the experiments. |
| Experiment Setup | Yes | For this task we implemented an ADMM solver from the work of Boyd et al. [4] with multiple stopping criterions including a fixed number of iterations of T = 1000. Our sample of n = 100 points are drawn from K = 5 identically-sized, perfectly discriminated non-isovolumic clusters of Gaussians... In (S1), for a given dimension space p = 500... signal-to-noise ratio increased from 1 to 15. In (S2) we impose a fixed signal to noise ratio and observe the algorithm s decay in performance as the space dimension p is increased from 102 to 105. Lloyd s K-means algorithm [13] with a thousand K-means++ initialization of [1]. |