SVM-Based Deep Stacking Networks

Authors: Jingyuan Wang, Kai Feng, Junjie Wu5273-5280

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

Reproducibility Variable Result LLM Response
Research Type Experimental Experimental results on both image and text data sets demonstrate the excellent performances of SVM-DSN compared with some competitive benchmark models.
Researcher Affiliation Academia MOE Engineering Research Center of Advanced Computer Application Technology, School of Computer Science Engineering, Beihang University, Beijing 100191, China Beijing Key Laboratory of Emergency Support Simulation Technologies for City Operations, School of Economics and Management, Beihang University, Beijing 100191, China Beijing Advanced Innovation Center for Big Data and Brain Computing, Beihang University, Beijing 100191, China Email: {jywang, fengkai, wujj}@buaa.edu.cn
Pseudocode Yes Algorithm 1 gives the pseudocodes of BLT.
Open Source Code No The paper does not provide an explicit statement or link for the open-source code of the described methodology.
Open Datasets Yes We first test the performance of SVM-DSN on the MNIST image classification database (Le Cun et al. 1998). ... We use the IMDB sentiment classification data set (Maas et al. 2011) in our experiment.
Dataset Splits Yes The MNIST database contains 60,000 handwritten digits images in the size of 28 28 for training and validation, and 10,000 images for testing.
Hardware Specification No The paper does not provide specific details regarding the hardware used for running its experiments.
Software Dependencies No The paper does not specify any software dependencies with their version numbers.
Experiment Setup Yes The SVM-DSN model used in the experiment consists of three layers two middle layers and an output layer. Both of the block in the two middle layers contains 20 base-SVM groups, and each group contains 10 base-SVMs, i.e., 200 base-SVMs one layer. ... In our experiment, we directly set the learning rate as a fix value η = 0.0005 to ensure the experiment fairness.