Rectangular Bounding Process

Authors: Xuhui Fan, Bin Li, Scott SIsson

NeurIPS 2018 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental The experimental results validate the merit of the RBP in rich yet parsimonious expressiveness compared to the state-of-the-art methods. The experimental results on a number of synthetic and real-world data sets demonstrate that the RBP can achieve parsimonious partitions with competitive performance compared to the state-of-the-art methods.
Researcher Affiliation Academia Xuhui Fan School of Mathematics & Statistics University of New South Wales xuhui.fan@unsw.edu.au Bin Li School of Computer Science Fudan University libin@fudan.edu.cn Scott A. Sisson School of Mathematics & Statistics University of New South Wales scott.sisson@unsw.edu.au
Pseudocode No The generative process for the RBP is described in numbered steps (e.g., '1. Sample the number of bounding boxes Kτ Poisson(τ QD d=1 1 + λL(d) );'), but it is presented as descriptive text rather than a formal, structured pseudocode block or algorithm.
Open Source Code No The paper mentions using existing code for comparison methods (e.g., 'for the MF, we directly use the existing code provided by the author1'), but it does not provide an explicit statement or link for the open-source code of their own Rectangular Bounding Process (RBP) implementation.
Open Datasets Yes Real-world data: We select several real-world data sets to compare the RBP-RT and the other state-of-the-art methods: Protein Structure [8] (N = 45, 730, D = 9), Naval Plants [7] (N = 11, 934, D = 16), Power Plants [34] (N = 9, 569, D = 4), Concrete [37] (N = 1, 030, D = 8), and Airfoil Self-Noise [8] (N = 1, 503, D = 8). Here, we first use PCA to select the 4 largest components and then normalize them so that they lie in the unit hypercube for ease of implementation. Dua Dheeru and EfiKarra Taniskidou. UCI machine learning repository, 2017.
Dataset Splits No The reported performance is averaged over 10 randomly selected hold-out test sets (Train:Test = 9:1). No specific validation set split is mentioned, only train and test.
Hardware Specification No No specific hardware details (such as GPU or CPU models, memory, or cloud instance types) used for running experiments are mentioned in the paper.
Software Dependencies No The paper mentions using 'scikit-learn toolbox [28]' and 'existing code provided by the author1' for comparison methods, but it does not specify version numbers for these or any other software dependencies needed to replicate the experiment.
Experiment Setup Yes To implement the RBP-TR we set the total number of iterations to 500, λ = 2 (i.e. the expected box length is 1/3) and τ = 1 (i.e. the expected number of bounding boxes is 90). In RBP-RM, we set λ = 0.99 and τ = 3, which leads to an expectation of 12 boxes.