Randomized Block Cubic Newton Method

Authors: Nikita Doikov, Peter Richtarik, University Edinburgh

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

Reproducibility Variable Result LLM Response
Research Type Experimental We establish O(1/ϵ), O(1/ϵ) and O(log(1/ϵ)) rates under different assumptions on the component functions. Lastly, we show numerically that our method outperforms the state of the art on a variety of machine learning problems, including cubically regularized leastsquares, logistic regression with constraints, and Poisson regression. Finally, our numerical experiments on synthetic and real datasets are described in Section 8.
Researcher Affiliation Academia 1National Research University Higher School of Economics, Samsung-HSE Laboratory, Moscow, Russia 2King Abdullah University of Science and Technology, Thuwal, Saudi Arabia 3University of Edinburgh, Edinburgh, United Kingdom 4Moscow Institute of Physics and Technology, Dolgoprudny, Russia.
Pseudocode Yes Algorithm 1 RBCN: Randomized Block Cubic Newton; Algorithm 2 Stochastic Dual Cubic Newton Ascent (SDCNA)
Open Source Code No The paper does not provide an unambiguous statement or a direct link to the open-source code for the methodology described.
Open Datasets No The paper mentions datasets like "leukemia" and "duke breast-cancer" but does not provide concrete access information (link, DOI, repository name, or formal citation with authors/year) for them to be considered publicly available according to the strict criteria.
Dataset Splits No The paper does not provide specific details regarding training, validation, or test dataset splits (e.g., percentages, sample counts, or explicit references to predefined splits).
Hardware Specification No The paper does not explicitly describe the specific hardware used for running its experiments, such as GPU/CPU models, memory, or detailed computer specifications.
Software Dependencies No The paper does not provide specific software dependencies or library versions needed to replicate the experiment.
Experiment Setup Yes Using middle-size blocks of coordinates on each step is the best choice in terms of total computational time; We see that using coordinate blocks of size 25 50 for the Cubic Newton outperforms all other cases of both methods in terms of total computational time; Comparison of Algorithm 2 (marked as Cubic) with SDNA and SDCA methods for minibatch sizes τ = 8, 32, 256, training Poisson regression.