Mind the Nuisance: Gaussian Process Classification using Privileged Noise

Authors: Daniel Hernández-lobato, Viktoriia Sharmanska, Kristian Kersting, Christoph H. Lampert, Novi Quadrianto

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

Reproducibility Variable Result LLM Response
Research Type Experimental Extensive experiments on public datasets show that the proposed GPC method using privileged noise, called GPC+, improves over a standard GPC without privileged knowledge, and also over the current state-of-the-art SVM-based method, SVM+.
Researcher Affiliation Academia Daniel Hern andez-Lobato Universidad Aut onoma de Madrid Madrid, Spain daniel.hernandez@uam.es Viktoriia Sharmanska IST Austria Klosterneuburg, Austria vsharman@ist.ac.at Kristian Kersting TU Dortmund Dortmund, Germany first.last@cs.tu-dortmund.de Christoph H. Lampert IST Austria Klosterneuburg, Austria chl@ist.ac.at Novi Quadrianto SMi Le CLi Ni C, University of Sussex Brighton, United Kingdom n.quadrianto@sussex.ac.uk
Pseudocode No The paper does not contain any structured pseudocode or algorithm blocks.
Open Source Code Yes An R language implementation of GPC+ using EP for approximate inference is found in the supplementary material.
Open Datasets Yes The data set was collected from a website that aggregates product data from a variety of e-commerce sources and includes both images and associated textual descriptions [20]. The dataset was collected by querying image search engines for each of the 50 animals categories which have complimentary high level descriptions of their semantic properties such as shape, colour, or habitat information among others [24]. All features are provided with the Aw A dataset2. 2http://attributes.kyb.tuebingen.mpg.de
Dataset Splits Yes We generated 6 binary classification tasks for each pair of the 4 classes with 200 samples for training, 200 samples for validation, and the rest of the samples for testing performance. ...we generated 45 binary classification tasks for each pair of the 10 classes with 200 samples for training, 200 samples for validation, and the rest of samples for testing the predictive performance.
Hardware Specification No No specific hardware details (e.g., CPU/GPU models, memory) used for running the experiments are provided in the paper.
Software Dependencies No The paper mentions 'An R language implementation' but does not specify the version of R or any other software dependencies with version numbers.
Experiment Setup Yes For all four methods, we used a squared exponential kernel with amplitude parameter θ and smoothness parameter l. For simplicity, we set θ = 1.0 in all cases. ...In GPC and GPC+, we used type II-maximum likelihood for finding all hyper-parameters.