Notice: The reproducibility variables underlying each score are classified using an automated LLM-based pipeline, validated against a manually labeled dataset. LLM-based classification introduces uncertainty and potential bias; scores should be interpreted as estimates. Full accuracy metrics and methodology are described in [1].

Memorization With Neural Nets: Going Beyond the Worst Case

Authors: Sjoerd Dirksen, Patrick Finke, Martin Genzel

JMLR 2024 | Venue PDF | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental We verify our theoretical result with numerical experiments and additionally investigate the effectiveness of the algorithm on MNIST and CIFAR-10.
Researcher Affiliation Collaboration Sjoerd Dirksen EMAIL Mathematical Institute Utrecht University 3584 CD Utrecht, Netherlands; Patrick Finke EMAIL Mathematical Institute Utrecht University 3584 CD Utrecht, Netherlands; Martin Genzel EMAIL Merantix Momentum GmbH 13355 Berlin, Germany
Pseudocode Yes Algorithm 1 Interpolation; Algorithm 2 Interpolation (experiments)
Open Source Code Yes Code is available at https://github.com/patrickfinke/memo. We use Python 3, Scikit-learn, and Num Py.
Open Datasets Yes we additionally investigate the effectiveness of the algorithm on MNIST and CIFAR-10. Examining binary classification subproblems of the MNIST data set (Le Cun et al., 1998)
Dataset Splits No The paper describes experiments on various datasets (Two Moons, MNIST, CIFAR-10) but does not specify explicit training/test/validation splits or methodologies used for splitting the data for its experiments.
Hardware Specification No The paper does not provide any specific details about the hardware (e.g., GPU/CPU models, memory) used for running the experiments.
Software Dependencies No Code is available at https://github.com/patrickfinke/memo. We use Python 3, Scikit-learn, and Num Py. The paper mentions software dependencies like Python 3, Scikit-learn, and NumPy, but it only provides a version number for Python. Specific version numbers are missing for Scikit-learn and NumPy.
Experiment Setup No The paper describes the algorithms and their performance on datasets but does not explicitly provide details about hyperparameters (e.g., learning rate, batch size, epochs, optimizers) or system-level training settings for the numerical experiments.