D-CIPHER: Discovery of Closed-form Partial Differential Equations

Authors: Krzysztof Kacprzyk, Zhaozhi Qian, Mihaela van der Schaar

NeurIPS 2023 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental We perform a series of synthetic experiments to show how well D-CIPHER is able to discover some well-known differential equations
Researcher Affiliation Collaboration Krzysztof Kacprzyk University of Cambridge kk751@cam.ac.uk Zhaozhi Qian University of Cambridge zq224@cam.ac.uk Mihaela van der Schaar University of Cambridge, The Alan Turing Institute mv472@cam.ac.uk
Pseudocode Yes The pseudocode of D-CIPHER is presented in Algorithm 1.
Open Source Code Yes All experiment code can be found at https://github.com/krzysztof-kacprzyk/D-CIPHER
Open Datasets No The paper states that fields for experiments were "generated by solving the equation" and observed fields by "sampling (t, x) ... and adding Gaussian noise" (Appendix E.5). This indicates synthetic data generation rather than the use of a publicly accessible, pre-existing dataset with a concrete link, DOI, or formal citation.
Dataset Splits No The paper describes how synthetic data is generated for experiments, mentioning "sampling grid" and "number of samples" (Appendix E.5, E.6). However, it does not specify traditional dataset splits such as exact percentages or sample counts for training, validation, and testing partitions of a fixed dataset.
Hardware Specification Yes The time is measured on a single computer with an Intel Core i5-6500 CPU (4 cores) and 16GB of RAM.
Software Dependencies No Table 7 lists "Software used and their licenses" (e.g., gplearn, cvxopt, sympy, scikit-learn, numpy, scipy, python, pysindy) but does not provide specific version numbers for these software components. For example, it lists "scikit-learn BSD 3-Clause" rather than "scikit-learn 1.2.3".
Experiment Setup Yes Table 6: Hyperparameters used in gplearn