Statistical Regimes and Runtime Prediction

Authors: Barry Hurley, Barry O'Sullivan

IJCAI 2015 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental Supported by a large-scale empirical study employing many years of industrial SAT Competition instances including repeated runs, we present statistical and empirical evidence that such a performance variation phenomenon necessitates a change in the evaluation of portfolio, runtime prediction, and automated configuration methods.
Researcher Affiliation Academia Barry Hurley and Barry O Sullivan Insight Centre for Data Analytics, University College Cork, Ireland
Pseudocode No The paper does not contain any pseudocode or clearly labeled algorithm blocks.
Open Source Code No The paper states that the dataset is available at a URL, but it does not provide an explicit statement or link for the open-source code of the methodology described in the paper.
Open Datasets Yes The dataset is available at http://ucc.insight-centre.org/bhurley/
Dataset Splits Yes We use a standard randomised 10-fold cross-validation, where the dataset is split in 10 folds.
Hardware Specification Yes Performance data was collected on a cluster of Intel Xeon E5430 2.66Ghz processors running Cent OS 6.
Software Dependencies Yes We use Mini Sat 2.0 [E en and S orensson, 2004] as the solver... running Cent OS 6.
Experiment Setup Yes For the predictions in this section we will use the adjusted logtransformed runtime, log10(1 + runtime). ... We use a random forest with the same parameters as [Hutter et al., 2014], that is using 10 regression trees, using half of the variables as split variables at each node (perc = 0.5), a maximum of 5 data-points in leaf nodes (nmin = 5).