k-variates++: more pluses in the k-means++

Authors: Richard Nock, Raphael Canyasse, Roksana Boreli, Frank Nielsen

ICML 2016 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental We demonstrate the applicability of our analysis via experimental evaluation on several domains and settings, displaying competitive performances vs state of the art. Section 5 presents experimental results.
Researcher Affiliation Collaboration Data61, Sydney, Australia & { The Australian National University; The University of New South Wales} Ecole Polytechnique, Palaiseau, France & { The Technion, Haifa, Israel; Sony CS Labs, Inc., Tokyo, Japan}
Pseudocode Yes Algorithm 0 k-variates++ ... Algorithm 1 Dk-means++ ... Algorithm 2 Sk-means++ ... Algorithm 3 OLk-means++
Open Source Code No The paper refers to a Supplementary Information (Nock et al., 2016a) for proofs and extensive experiments, but does not provide a direct link to any open-source code repository for the methodology.
Open Datasets No The paper mentions datasets like 'Life Sci', 'Image', 'Europe Diff' in Table 2 and 'synthetic data', but does not provide concrete access information (e.g., specific links, DOIs, or citations with authors/year) for public availability.
Dataset Splits No The paper does not explicitly provide training/validation/test dataset splits, percentages, or cross-validation details needed to reproduce the experiment.
Hardware Specification No The paper does not explicitly describe the specific hardware (e.g., GPU/CPU models, memory details) used to run its experiments.
Software Dependencies No The paper does not provide specific version numbers for ancillary software components used in the experiments.
Experiment Setup Yes Parameters are in line with (Bahmani et al., 2012). ... We let ϵ = 1 in our experiments. ... each peer s initial data consists of points uniformly sampled in a random hyperrectangle in a space of d = 50 (expected number of peers points mi = 500, i). We sample peers until a total of m 20000 point is sampled.