L2-Nonexpansive Neural Networks

Authors: Haifeng Qian, Mark N. Wegman

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

Reproducibility Variable Result LLM Response
Research Type Experimental Experiments are divided into three groups to study different properties of L2NNNs.
Researcher Affiliation Industry Haifeng Qian & Mark N. Wegman IBM Research Yorktown Heights, NY 10598, USA qianhaifeng,wegman@us.ibm.com
Pseudocode No Not found. The paper describes its methods in prose and mathematical formulas, but does not include any explicit pseudocode or algorithm blocks.
Open Source Code Yes Our MNIST and CIFAR-10 classifiers are available at http://researcher.watson.ibm.com/group/9298
Open Datasets Yes With MNIST and CIFAR-10 classifiers
Dataset Splits Yes In early-stopping runs, 5000 training images are withheld as validation set and training stops when loss on validation set stops decreasing.
Hardware Specification No Not found. The paper does not provide specific hardware details (e.g., exact GPU/CPU models, processor types, or memory amounts) used for running its experiments.
Software Dependencies No Not found. The paper does not provide specific ancillary software details (e.g., library or solver names with version numbers) needed to replicate the experiment.
Experiment Setup Yes We use a loss function with three terms, with trade-off hyperparameters γ and ω: L = La + γ Lb + ω Lc (3)