Compressed Spectral Regression for Efficient Nonlinear Dimensionality Reduction

Authors: Deng Cai

IJCAI 2015 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental In this section, we conduct the clustering experiment on the MNIST (handwritten digits image) data set to demonstrate the effectiveness of the proposed Compressed Spectral Regression (CSR) approach. Each image is represented as a 784 dimensional vector. It has 60000 training images and 10000 test images.
Researcher Affiliation Academia Deng Cai The State Key Lab of CAD&CG, College of Computer Science, Zhejiang University, China dengcai@cad.zju.edu.cn
Pseudocode Yes Algorithm 1 Compressed Spectral Regerssion
Open Source Code No The paper does not provide concrete access to source code for the methodology described.
Open Datasets Yes In this section, we conduct the clustering experiment on the MNIST (handwritten digits image) data set to demonstrate the effectiveness of the proposed Compressed Spectral Regression (CSR) approach. Each image is represented as a 784 dimensional vector. It has 60000 training images and 10000 test images.
Dataset Splits No The paper specifies training and test sets but does not explicitly mention a separate validation set split, percentages, or sample counts.
Hardware Specification Yes All the codes in the experiments are implemented in MATLAB R2011b and run on a Windows 7 machine with 3.40 GHz i72600K CPU, 16GB main memory.
Software Dependencies Yes All the codes in the experiments are implemented in MATLAB R2011b and run on a Windows 7 machine with 3.40 GHz i72600K CPU, 16GB main memory.
Experiment Setup Yes For CSR, we empirically set the parameters l = 1000 (# landmarks), t = 5 (# iterations in k-means), r = 5 (# nearest landmarks) and α = 0.01 (regression regularization). We use the same Gaussian kernel for all the kernel methods and our CSR in Eq. (7). The kernel width parameter σ is empirically set to be 10.