L1-regularized Neural Networks are Improperly Learnable in Polynomial Time

Authors: Yuchen Zhang, Jason D. Lee, Michael I. Jordan

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

Reproducibility Variable Result LLM Response
Research Type Experimental In this section, we compare the proposed algorithm with several baseline algorithms on the MNIST digit recognition task.
Researcher Affiliation Academia Yuchen Zhang YUCZHANG@EECS.BERKELEY.EDU Jason D. Lee JASONDLEE@BERKELEY.EDU Michael I. Jordan JORDAN@CS.BERKELEY.EDU Department of EECS, University of California, Berkeley, CA 94720 USA
Pseudocode Yes Algorithm 1 Recursive Kernel Method Input: Feature-label pairs {(xi, yi)}n i=1; Loss function ℓ: R R R; Number of hidden layers k; Regularization coefficient B.
Open Source Code No The paper does not provide any specific links or explicit statements about releasing the source code for the proposed recursive kernel method.
Open Datasets Yes We use the MNIST handwritten digits dataset and three variations of it. See Figure 2 for the description of these datasets and several exemplary images. Variations on the MNIST digits. http://www.iro. umontreal.ca/ lisa/twiki/bin/view. cgi/Public/Mnist Variations.
Dataset Splits Yes For all datasets, we use 10,000 images for training, 2,000 images for validation and 50,000 images for testing.
Hardware Specification No The paper does not provide any specific hardware details such as GPU models, CPU types, or cloud instance specifications used for running the experiments.
Software Dependencies No The paper does not list specific software dependencies with their version numbers, such as programming languages, libraries, or frameworks used for implementation.
Experiment Setup Yes For the recursive kernel method, we train one-vs-all SVM classifiers with Algorithm 1. The hyper-parameters are given by k {1, 4} and B = 100. All images are preprocessed by the following steps: deskewing, centering and normalization.