Fine-grained Generalization Analysis of Vector-Valued Learning
Authors: Liang Wu, Antoine Ledent, Yunwen Lei, Marius Kloft10338-10346
AAAI 2021 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | In this section, we present experimental results to verify our theoretical analysis. We consider a specific vectorvalued learning problem called multinomial logistic regression, where the aim is to predict a class label for each training example. We apply SGD (7) to solve the optimization problem with the objective function... We consider four real-world datasets available from the LIBSVM homepage (Chang and Lin 2011)... We compute both the training error FS(w T ) and testing error F(w T ) of the output model w T (last iterate). In this way we get a training error and a testing error for each considered sample size. We then plot the relative behavior of these errors versus the number of training examples in Figure 2. |
| Researcher Affiliation | Academia | Liang Wu 1, Antoine Ledent 2, Yunwen Lei 3, 2 and Marius Kloft 2 1Center of Statistical Research, School of Statistics, Southwestern University of Finance and Economics, Chengdu 611130, China 2Department of Computer Science, TU Kaiserslautern, 67653 Kaiserslautern, Germany 3School of Computer Science, University of Birmingham, Birmingham B15 2TT, United Kingdom |
| Pseudocode | No | The paper describes algorithms but does not include structured pseudocode or algorithm blocks with explicit labels like 'Algorithm' or 'Pseudocode'. |
| Open Source Code | No | The paper does not provide any statements about releasing code for the described methodology or links to a source code repository. |
| Open Datasets | Yes | We consider four real-world datasets available from the LIBSVM homepage (Chang and Lin 2011), whose information is summarized in the ar Xiv version. |
| Dataset Splits | No | For the first experiment, we randomly use 80% of data for training and reserve the remaining 20% for testing. The paper does not explicitly mention a separate validation split. |
| Hardware Specification | No | The paper does not provide any specific details about the hardware (e.g., GPU/CPU models, memory) used for running the experiments. |
| Software Dependencies | No | The paper mentions applying SGD and using datasets from LIBSVM homepage but does not specify version numbers for any software dependencies, libraries, or frameworks used in the implementation. |
| Experiment Setup | Yes | We set the initial point w = 0 and the step size ηt = 1/(λt + 1). We repeat experiments 50 times and report the average as well as standard deviation of the experimental results. ... where λ = 0.01. |