Memory and Computation Efficient PCA via Very Sparse Random Projections
Authors: Farhad Pourkamali Anaraki, Shannon Hughes
ICML 2014 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | We present experimental results demonstrating that this approach allows for simultaneously achieving a substantial reduction of the computational complexity and memory/storage space, with little loss in accuracy, particularly for very high-dimensional data. |
| Researcher Affiliation | Academia | Farhad Pourkamali-Anaraki FARHAD.POURKAMALI@COLORADO.EDU Shannon M. Hughes SHANNON.HUGHES@COLORADO.EDU Department of Electrical, Computer, and Energy Engineering, University of Colorado at Boulder, CO, 80309, USA |
| Pseudocode | No | The paper does not contain any pseudocode or algorithm blocks. |
| Open Source Code | No | The paper does not provide any explicit statements or links indicating the release of its source code. |
| Open Datasets | Yes | Finally, we consider the MNIST dataset to see a real-world application outside the spiked covariance model. This dataset contains 70,000 samples of handwritten digits, which we have resized to 40 40 pixels. Hence, we have 70,000 samples in R1600. |
| Dataset Splits | No | The paper does not specify exact training, validation, or test dataset splits. It mentions the total number of samples for MNIST but no partitioning. |
| Hardware Specification | No | The paper does not explicitly describe the specific hardware used to run its experiments. |
| Software Dependencies | No | The paper mentions 'MATLAB s svds' but does not specify a version number for MATLAB or any other software dependencies with their versions. |
| Experiment Setup | No | The paper discusses parameters like SNR (Signal-to-Noise Ratio), measurement ratio (m/p), and compression factor (γ) that are integral to its method, but it does not provide specific hyperparameter values or system-level training settings typically found in experimental setups (e.g., learning rates, batch sizes, optimizers). |