Low-Complexity Nonparametric Bayesian Online Prediction with Universal Guarantees
Authors: Alix LHERITIER, Frederic Cazals
NeurIPS 2019 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | Experiments on challenging datasets show the computational and statistical efficiency of our algorithm in comparison to standard and state-of-the-art methods. |
| Researcher Affiliation | Collaboration | Alix Lhéritier Amadeus SAS F-06902 Sophia-Antipolis, France alix.lheritier@amadeus.com Frédéric Cazals Université Côte d Azur Inria F-06902 Sophia-Antipolis, France frederic.cazals@inria.fr |
| Pseudocode | No | Section 5 'Online algorithm' describes the algorithm's steps in prose, stating 'The steps of our algorithm are the same as those of [33, Algorithm 1]'. However, it does not present a structured pseudocode block or algorithm box within the paper itself. |
| Open Source Code | Yes | Python code and data used for the experiments are available at https: //github.com/alherit/kd-switch. |
| Open Datasets | Yes | We use the following datasets, detailed in Appendix B.1: (L-i) A 2D dataset consists of two Gaussian Mixtures spanning three different scales. (L-ii) A dataset in dimension d = 784 composed of both real MNIST digits, as well as digits generated by a Generative Adversarial Network [24] trained on the MNIST dataset. (L-iii) The Higgs dataset [20], the goal being to distinguish the signature of processes producing Higgs bosons. (L-iv) The Breast Cancer Wisconsin (Diagnostic) Data Set [20] dimension d = 30. |
| Dataset Splits | No | The paper describes how data is fed to online predictors and mentions a 'train-test paradigm' in the context of comparing with other methods (e.g., 'the train-test paradigm as opposed to KDS-seq which automatically detects the pertinent scales'), but it does not specify explicit train/validation/test splits, percentages, or cross-validation details for its own experiments. |
| Hardware Specification | Yes | Experiments were carried out on a machine running Debian 3.16, equipped with two Intel(R) Xeon(R) E5-2667 v2 @ 3.30GHz processors and 62 GB of RAM. |
| Software Dependencies | No | The paper mentions 'Python code' and states 'Our implementation uses the scikitlearn Gaussian Process Classifier [23]', but it does not specify version numbers for Python or scikit-learn. |
| Experiment Setup | Yes | We compare the performance of our online predictors Pkds and Pkdw (see Rmk. 2) with a number of trees J {1, 50}... The Bayesian Mixture of Gaussian Processes Classifiers (gp) with RBF kernel width σ {24i}i= 5...7... The significance level is set to α = .01 in all the cases. |