Adaptive Sampling for Discovery

Authors: Ziping Xu, Eunjae Shim, Ambuj Tewari, Paul Zimmerman

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

Reproducibility Variable Result LLM Response
Research Type Experimental The benefits of IDS are shown in both simulation experiments and real-data experiments for discovering chemical reaction conditions. We start with simulation studies on the three problems: (generalized) linear model, low-rank matrix and graph model.
Researcher Affiliation Academia Ziping Xu Department of Statistics University of Michigan zipingxu@umich.edu Eunjae Shim Department of Chemistry University of Michigan eunjae@umich.edu Ambuj Tewari Department of Statistics University of Michigan tewaria@umich.edu Paul Zimmerman Department of Chemistry University of Michigan paulzim@umich.edu
Pseudocode Yes Algorithm 1 IDS (Information-Directed Sampling) for Discovery
Open Source Code Yes Did you include the code, data, and instructions needed to reproduce the main experimental results (either in the supplemental material or as a URL)? [Yes] The codes and dataset are in the supplemental material.
Open Datasets Yes The first dataset, named Photoredox Nickel Dual-Catalysis (PNDC) [11], comes from a reaction that involves two types of catalysts... The second C-N Cross-Coupling with Isoxazoles (CNCCI) dataset [2] comes from a collection of a cross-coupling reactions...
Dataset Splits No The paper states 'They [training details, e.g. data splits] are provided along with code' in the checklist, but the main text does not explicitly detail the training/validation/test dataset splits, percentages, or specific cross-validation setup.
Hardware Specification No The checklist states 'Did you include the total amount of compute and the type of resources used...? [Yes] They are given in Appendix H'. However, Appendix H within the provided paper content does not contain specific hardware details such as GPU/CPU models or cloud provider specifications.
Software Dependencies No The paper does not explicitly list specific software dependencies with version numbers (e.g., library names with versions) in its main text or appendices.
Experiment Setup Yes The paper mentions specific experimental parameters for simulation studies: 'We study the effects of d = 20, 50, 100' and 'We tested σ2 ϵ = 0.1, 1.0, 10'. For real-data experiments, it states, 'We tuned hyperparameters for all three algorithms. Details are given in Appendix H.'