Sparsified SGD with Memory

Authors: Sebastian U. Stich, Jean-Baptiste Cordonnier, Martin Jaggi

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

Reproducibility Variable Result LLM Response
Research Type Experimental We present numerical experiments to illustrate the theoretical findings and the good scalability for distributed applications.
Researcher Affiliation Academia Machine Learning and Optimization Laboratory (MLO) EPFL, Switzerland
Pseudocode Yes The pseudocode is given in Algorithm 1.
Open Source Code Yes Our code is open-source and publicly available at github.com/epfml/sparsified SGD.
Open Datasets Yes Datasets. We consider a dense dataset, epsilon [35], as well as a sparse dataset, RCV1 [20] where we train on the larger test set.
Dataset Splits No The paper mentions using datasets for training but does not provide specific details on training/validation/test splits, such as percentages, sample counts, or explicit splitting methodology.
Hardware Specification Yes Experiments were run on an Ubuntu 18.04 machine with a 24 cores processor Intel Xeon CPU E5-2680 v3 @ 2.50GHz.
Software Dependencies No The paper mentions 'Python3', 'numpy library', and 'scikit-learn' but does not provide specific version numbers for the libraries.
Experiment Setup Yes We study the convergence of the method using the stepsizes ηt = γ/(λ(t+a)) and hyperparameters γ and a set as in Table 2.