Stochastic Cubic Regularization for Fast Nonconvex Optimization

Authors: Nilesh Tripuraneni, Mitchell Stern, Chi Jin, Jeffrey Regier, Michael I. Jordan

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

Reproducibility Variable Result LLM Response
Research Type Experimental We also empirically show that the stochastic cubic-regularized Newton method proposed in this paper performs favorably on both synthetic and real non-convex problems relative to state-of-the-art optimization methods.
Researcher Affiliation Academia Nilesh Tripuraneni Mitchell Stern Chi Jin Jeffrey Regier Michael I. Jordan {nilesh_tripuraneni,mitchell,chijin,regier}@berkeley.edu jordan@cs.berkeley.edu University of California, Berkeley
Pseudocode Yes Algorithm 1 Stochastic Cubic Regularization (Meta-algorithm) Input: mini-batch sizes n1, n2, initialization x0, number of iterations Tout, and final tolerance .
Open Source Code No The paper does not provide any links or explicit statements about the public release of its source code.
Open Datasets Yes training a deep autoencoder on MNIST [Le Cun and Cortes, 2010].
Dataset Splits No The paper mentions training on MNIST but does not specify the train/validation/test split percentages or sample counts for each partition.
Hardware Specification No The paper mentions software used for implementation (TensorFlow) but does not provide specific details about the hardware (e.g., GPU or CPU models) used for running the experiments.
Software Dependencies No The paper mentions TensorFlow as the implementation framework but does not provide specific version numbers for TensorFlow or any other software dependencies.
Experiment Setup Yes The batch sizes and learning rates for each method are tuned separately to ensure a fair comparison; see Appendix D for details.