Towards an optimal stochastic alternating direction method of multipliers
Authors: Samaneh Azadi, Suvrit Sra
ICML 2014 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | We present several experiments with our new methods; the results indicate improved performance over competing ADMM methods. In this section we present experiments that illustrate performance of our SADMM variants. |
| Researcher Affiliation | Academia | Samaneh Azadi SAZADI156@GMAIL.COM UC Berkeley, Berkeley, CA School of ECE, Shiraz University, Shiraz, Iran Suvrit Sra SUVRIT@TUEBINGEN.MPG.DE Carnegie Mellon University, Pittsburgh Max Planck Institute for Intelligent Systems, T ubingen, Germany |
| Pseudocode | Yes | Algorithm 1: ADMM, Algorithm 2: Stochastic ADMM (strongly convex), Algorithm 3: SADMM for smooth f(x) |
| Open Source Code | No | The paper does not provide an explicit statement or link to the open-source code for the methodology described in the paper. It only mentions the use of a 'freely available implementation of the LBFGS-B method' which is a third-party tool. |
| Open Datasets | Yes | We compare these methods on a version of the well-known 20newsgroups dataset3. This dataset consists of binary occurrence data of 100 words for 16,242 instances, and the samples are labeled into four categories for which one can do classification by one-vs rest scheme multiclass classification. (3) Obtained from http://www.cs.nyu.edu/ roweis/data.html and We used the dataset adult5 which contains 123 dimensional feature vectors. (5) Obtained from the LIBSVM datasets webpage. |
| Dataset Splits | No | The paper mentions '20% of samples' for test data but does not explicitly provide details about a validation dataset split or percentages for training, validation, and test splits. |
| Hardware Specification | No | The paper does not explicitly describe the hardware used to run its experiments (e.g., specific CPU/GPU models, memory, or cloud instance details). |
| Software Dependencies | No | The paper mentions using 'the well-known freely available implementation of the LBFGS-B method' for solving a dual problem, but it does not provide specific version numbers for this or any other key software components used in the experiments. |
| Experiment Setup | Yes | We note that for all all experiments, we set the AL parameter β = 1, as also done in Ouyang et al. (2013). Hyperparameter C is set to 0.01. Moreover, we used mini-batches of size 10 for each iteration. Step size equal to 1/k is used for the SGD and proximal SGD methods. |