Simple Classification Using Binary Data
Authors: Deanna Needell, Rayan Saab, Tina Woolf
JMLR 2018 | Venue PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | We illustrate the utility of the proposed approach through stylized and realistic numerical experiments, and provide a theoretical analysis for a simple case. We hope that our framework and analysis will serve as a foundation for studying similar types of approaches. |
| Researcher Affiliation | Academia | Deanna Needell EMAIL Department of Mathematics 520 Portola Plaza, University of California, Los Angeles, CA 90095 Rayan Saab EMAIL Department of Mathematics 9500 Gilman Drive, University of California, La Jolla, CA 92093 Tina Woolf EMAIL Institute of Mathematical Sciences 150 E. 10th Street, Claremont Graduate University, Claremont CA 91711 |
| Pseudocode | Yes | Algorithm 1 Training Algorithm 2 Classification |
| Open Source Code | No | The paper does not provide an explicit statement or link to open-source code for the described methodology. The license information provided is for the paper itself, not for any accompanying code. |
| Open Datasets | Yes | In this section, we apply Algorithms 1 and 2 to the MNIST (Le Cun, 2018) dataset, which is a benchmark dataset of images of handwritten digits, each with 28 28 pixels. ... Our last experiment considers facial recognition using the extended Yale B dataset (Cai et al., 2007b,a, 2006; He et al., 2005). |
| Dataset Splits | Yes | In total, the dataset has 60, 000 training examples and 10, 000 testing examples. ... We utilize 1, 000, 3, 000, and 5, 000 training points per digit class, and perform classification with 800 test images per class. ... We select four individuals from the dataset, and randomly select images with different illuminations to be included in the training and testing sets ... p {20, 40, 60} with equally sized training data sets for each class, and classify 30 images per class. |
| Hardware Specification | No | The paper does not explicitly describe the specific hardware (e.g., GPU/CPU models, memory) used to run the experiments. It only discusses the general concept of hardware implementations in the introduction, not the experimental setup. |
| Software Dependencies | No | The paper does not provide specific software dependencies (e.g., library names with version numbers) used for implementing the described methodology. It mentions third-party tools like Caffe in the related work section, but not in the context of their own implementation with versions. |
| Experiment Setup | Yes | For each choice of m {5, 7, 9, 11, 13, 15, 17, 19} and p {75, 150, 225} with equally sized training data sets for each class (that is, each class is tested with either 25, 50, or 75 training points), we execute Algorithms 1 and 2 with a single level and 30 trials of generating A. We perform classification of 50 test points per group... In Figure 2, ... using four levels and m {10, 30, 50, 70, 90, 110, 130}. |