Learning Deep ℓ0Encoders

Authors: Zhangyang Wang, Qing Ling, Thomas Huang

AAAI 2016 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental Numerical results demonstrate the impressive performances of the proposed encoders.
Researcher Affiliation Academia Beckman Institute, University of Illinois at Urbana-Champaign, Urbana, IL 61801, USA Department of Automation, University of Science and Technology of China, Hefei, 230027, China
Pseudocode No The paper includes mathematical equations and block diagrams but no explicitly labeled 'Pseudocode' or 'Algorithm' sections, nor any structured code-like procedures.
Open Source Code No The paper does not provide any specific links, statements, or mentions of source code availability for the described methodology.
Open Datasets Yes The first 60,000 samples of the MNIST dataset are used for training and the last 10,000 for testing. ... We evaluate our methods on the MNIST dataset, and the AVIRIS Indiana Pines hyperspectral image dataset (see (Wang, Nasrabadi, and Huang 2015) for details). ... For clustering, we evaluate our methods on the COIL 20 and the CMU PIE dataset (Sim, Baker, and Bsat 2002).
Dataset Splits No The paper specifies 'The first 60,000 samples of the MNIST dataset are used for training and the last 10,000 for testing,' but does not explicitly mention a validation split, its size, or how it was used.
Hardware Specification Yes In practice, given that the model is well initialized, the training takes approximately 1 hour on the MNIST dataset, on a workstation with 12 Intel Xeon 2.67GHz CPUs and 1 GTX680 GPU.
Software Dependencies No The paper states 'implemented with the CUDA Conv Net package (Krizhevsky, Sutskever, and Hinton 2012)' but does not provide a specific version number for this package or any other software dependencies.
Experiment Setup Yes We use a constant learning rate of 0.01 with no momentum, and a batch size of 128.