Resolving the data ambiguity for periodic crystals

Authors: Daniel Widdowson, Vitaliy Kurlin

NeurIPS 2022 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental The strength of PDD was experimentally checked for all 660K+ periodic crystals in the world s largest collection CSD (Cambridge Structural Database). Despite the CSD being curated to contain only real and distinct structures [30], the new invariants identified several pairs of duplicates. All the underlying publications are now being investigated for data integrity by five journals, see section 6.
Researcher Affiliation Academia Daniel E Widdowson Department of Computer Science University of Liverpool, Liverpool, L69 3BX D.E.Widdowson@liverpool.ac.uk Vitaliy A Kurlin Department of Computer Science University of Liverpool, Liverpool, L69 3BX vitaliy.kurlin@gmail.com
Pseudocode No The paper describes the algorithm and its complexity (Theorem 5.1) but does not include structured pseudocode or an algorithm block.
Open Source Code Yes Yes, in the supplementary materials [66].
Open Datasets Yes The strength of PDD was experimentally checked for all 660K+ periodic crystals in the world s largest collection CSD (Cambridge Structural Database).
Dataset Splits No The paper states 'Yes' to 'Did you specify all the training details (e.g., data splits, hyperparameters, how they were chosen)?' in the NeurIPS checklist, but only specifies 'The PDD algorithm used one parameter k = 100 (the number of atomic neighbors) in addition to a typical input (Crystallographic Information File).' It does not provide specific data splits like percentages or counts for training, validation, or testing.
Hardware Specification Yes 200B+ pairwise comparisons of all periodic structures in the world s largest collection (Cambridge Structural Database) of existing materials over two days on a modest desktop. AMD Ryzen 5 5600X (6-core) @4.60Ghz, 32GB DDR4 RAM @3600 Mhz.
Software Dependencies No The paper does not provide specific software dependencies (e.g., library names with version numbers) needed to replicate the experiment.
Experiment Setup Yes The PDD algorithm used one parameter k = 100 (the number of atomic neighbors) in addition to a typical input (Crystallographic Information File).