One-shot Distributed Ridge Regression in High Dimensions

Authors: Yue Sheng, Edgar Dobriban

ICML 2020 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental Our results are supported by simulations and real data analysis. ... Section 4 contains experiments on real data. ... We confirm these results in detailed simulation studies and on an empirical data example, using the Million Song Dataset.
Researcher Affiliation Academia 1Wharton Statistics Department, University of Pennsylvania, Philadelphia, PA, USA 2Graduate Group in Applied Mathematics and Computational Science, University of Pennsylvania, Philadelphia, PA, USA.
Pseudocode Yes Algorithm 1: Optimally weighted distributed ridge regression
Open Source Code No No statement or link providing concrete access to source code for the methodology was found.
Open Datasets Yes Million Song Year Prediction Dataset (MSD) (Bertin-Mahieux et al., 2011). ... We download the dataset from the UC Irvine Machine Learning Repository.
Dataset Splits Yes For each experiment, we randomly choose ntrain = 10,000 samples from the training set and ntest = 1,000 samples from the test set.
Hardware Specification No No specific hardware details (e.g., GPU/CPU models, memory) used for experiments are mentioned.
Software Dependencies No No specific software versions (e.g., Python 3.8, PyTorch 1.9) are mentioned.
Experiment Setup Yes We choose the number of machines to be k = 1, 10, 20, 50, 100, 500, 1, 000, and we distribute the data evenly across the k machines. ... We repeat the experiment T = 100 times... ... The estimator using only a fraction 1/k of the data, which is just one of the local estimators. For this estimator, we choose the tuning parameter λ = kp/(ntrain ˆα2).