IDEAL: Inexact DEcentralized Accelerated Augmented Lagrangian Method

Authors: Yossi Arjevani, Joan Bruna, Bugra Can, Mert Gurbuzbalaban, Stefanie Jegelka, Hongzhou Lin

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

Reproducibility Variable Result LLM Response
Research Type Experimental We provide experimental results that demonstrate the effectiveness of the proposed algorithm on highly ill-conditioned problems.
Researcher Affiliation Academia Yossi Arjevani NYU yossia@nyu.eduJoan Bruna NYU bruna@cims.nyu.eduBugra Can Rutgers University bc600@scarletmail.rutgers.eduMert Gürbüzbalaban Rutgers University mg1366@rutgers.eduStefanie Jegelka MIT stefje@csail.mit.eduHongzhou Lin MIT hongzhou@mit.edu
Pseudocode Yes Algorithm 1 Decentralized Augmented Lagrangian framework; Algorithm 2 Accelerated Decentralized Augmented Lagrangian framework; Algorithm 3 IDEAL: Inexact Acc-Decentralized Augmented Lagrangian framework
Open Source Code No The paper does not contain any explicit statements or links indicating that the source code for the proposed methodology is publicly available.
Open Datasets Yes To facilitate a simple comparison between existing state-of-the-art algorithms, we consider an ℓ2-regularized logistic regression task over two classes of the MNIST [24]/CIFAR-10 [23] benchmark datasets. ... [24] Y. Le Cun, C. Cortes, and C. Burges. Mnist handwritten digit database. ATT Labs [Online], 2, 2010. URL http://yann.lecun.com/exdb/mnist.
Dataset Splits No The paper mentions using MNIST and CIFAR-10 datasets but does not provide specific details on how these datasets were split into training, validation, or test sets (e.g., percentages, counts, or explicit references to standard splits).
Hardware Specification No The paper does not provide specific details about the hardware used to run the experiments, such as GPU/CPU models, processor types, or memory.
Software Dependencies No The paper mentions using logistic regression and convolutional kernel networks but does not provide specific software dependencies (e.g., programming languages, libraries, or frameworks) with version numbers.
Experiment Setup Yes We set the inner iteration counter to be Tk = 100 for all algorithms, and use the theoretical stepsize schedule. The decentralized environment is modelled in a synthetic setting, where the communication time is steady and no latency is encountered. To demonstrate the effect of the underlying network architecture, we consider: a) a circular graph, where the agents form a cycle; b) a Barbell graph, where the agents are split into two complete subgraphs, connected by a single bridge (shown in Figure 2 in the appendix).