Non-Ergodic Convergence Analysis of Heavy-Ball Algorithms

Authors: Tao Sun, Penghang Yin, Dongsheng Li, Chun Huang, Lei Guan, Hao Jiang5033-5040

AAAI 2019 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental Experimental Results We report the numerical simulations of Heavy-ball method applied to the linear regression problem... As illustrated by Figure 1, larger β leads to faster convergence...
Researcher Affiliation Academia 1College of Computer, National University of Defense Technology, Changsha, Hunan, China. 2Department of Mathematics, University of California, Los Angeles, USA.
Pseudocode No The paper describes algorithms using mathematical equations and textual explanations, but does not provide structured pseudocode or algorithm blocks.
Open Source Code No The paper does not provide any statement or link indicating the release of open-source code for the described methodology.
Open Datasets No The data Ai and yi were generated by the Gaussian random and Bernoulli random distributions, respectively.
Dataset Splits No The paper does not provide specific percentages, sample counts, or citations for dataset splits (training, validation, or test) needed for reproduction.
Hardware Specification Yes All experiments were performed using MATLAB on an desktop with an Intel 3.4 GHz CPU.
Software Dependencies No The paper mentions 'MATLAB' but does not specify a version number or other software dependencies with their versions.
Experiment Setup Yes We fixed the stepsize as γ = 1 L in all numerical tests. For the stepsize, we need 2(1 βk) > 1, i.e., 0 βk < 0.5. Therefore, inertial parameters are set to βk β = 0, 0.1, 0.2, 0.3, 0.4. ... And we set n = 100 and m = 150. The data Ai and yi were generated by the Gaussian random and Bernoulli random distributions, respectively. The maximum number of iterations was set to 1000. For logistic regression, we set λ = 10 3.