Digital Discovery, 2023, 2,1269-1288
DOI: 10.1039/D3DD00023K, Paper

Open Access

  This article is licensed under a Creative Commons Attribution 3.0 Unported Licence.

Taylor J. Santaloci, William E. Meador, Austin M. Wallace, E. Michael Valencia, Blake N. Rogers, Jared H. Delcamp, Ryan C. Fortenberry
A readily accessible dye molecule with potential properties well-beyond the state-of-the-art for dye-sensitized solar cells is realized from extensive quantum chemical characterization of nearly 8000 stochastically-derived novel molecules.
The content of this RSS Feed (c) The Royal Society of Chemistry

Digital Discovery, 2023, 2,1334-1346
DOI: 10.1039/D2DD00149G, Paper

Open Access

  This article is licensed under a Creative Commons Attribution-NonCommercial 3.0 Unported Licence.

Siyu Isaac Parker Tian, Zekun Ren, Selvaraj Venkataraj, Yuanhang Cheng, Daniil Bash, Felipe Oviedo, J. Senthilnath, Vijila Chellappan, Yee-Fun Lim, Armin G. Aberle, Benjamin P. MacLeod, Fraser G. L. Parlane, Curtis P. Berlinguette, Qianxiao Li, Tonio Buonassisi, Zhe Liu
thicknessML predicts film thickness from reflection and transmission spectra. Transfer learning enables thickness prediction of different materials with good performance. Transfer learning also bridges the gap between simulation and experiment.
The content of this RSS Feed (c) The Royal Society of Chemistry

Digital Discovery, 2023, 2,1311-1325
DOI: 10.1039/D2DD00136E, Paper

Open Access

  This article is licensed under a Creative Commons Attribution-NonCommercial 3.0 Unported Licence.

Cayque Monteiro Castro Nascimento, Paloma Guimarães Moura, Andre Silva Pimentel
The local interpretable model-agnostic explanations method was used to interpret a machine learning model of toxicology generated by a neural network multitask classifier method.
The content of this RSS Feed (c) The Royal Society of Chemistry

Digital Discovery, 2023, 2,1354-1367
DOI: 10.1039/D3DD00060E, Paper

Open Access

  This article is licensed under a Creative Commons Attribution 3.0 Unported Licence.

Daniel R. Wong, David J. Logan, Santosh Hariharan, Robert Stanton, Djork-Arné Clevert, Andrew Kiruluta
Fluorescent-based microscopy screens carry a broad range of phenotypic information about how compounds affect cellular biology.
The content of this RSS Feed (c) The Royal Society of Chemistry

Digital Discovery, 2023, 2,1425-1435
DOI: 10.1039/D3DD00082F, Paper

Open Access

  This article is licensed under a Creative Commons Attribution 3.0 Unported Licence.

Eric S. Muckley, James E. Saal, Bryce Meredig, Christopher S. Roper, John H. Martin
On average, simple linear models perform equivalently to black box machine learning models on extrapolation tasks.
The content of this RSS Feed (c) The Royal Society of Chemistry

Digital Discovery, 2023, 2,1368-1379
DOI: 10.1039/D3DD00030C, Paper

Open Access

  This article is licensed under a Creative Commons Attribution 3.0 Unported Licence.

Noah Hoffmann, Jonathan Schmidt, Silvana Botti, Miguel A. L. Marques
Pretraining on large, lower-fidelity datasets enables extremely effective training of graph neural networks on smaller, high-fidelity datasets.
The content of this RSS Feed (c) The Royal Society of Chemistry

Digital Discovery, 2023, 2,1414-1424
DOI: 10.1039/D3DD00071K, Paper

Open Access

  This article is licensed under a Creative Commons Attribution 3.0 Unported Licence.

Henrik Schopmans, Patrick Reiser, Pascal Friederich
We used synthetically generated crystals to train ResNet-like models to enhance the prediction of space groups from ICSD powder X-ray diffractograms. The results show improved generalization to unseen structure types compared to previous approaches.
The content of this RSS Feed (c) The Royal Society of Chemistry

Digital Discovery, 2023, 2,1404-1413
DOI: 10.1039/D3DD00114H, Paper

Open Access

  This article is licensed under a Creative Commons Attribution-NonCommercial 3.0 Unported Licence.

Weiyi Gong, Tao Sun, Hexin Bai, Shah Tanvir ur Rahman Chowdhury, Peng Chu, Anoj Aryal, Jie Yu, Haibin Ling, John P. Perdew, Qimin Yan
We demonstrate that contrastive representation learning is a computationally efficient and flexible method to incorporate physical constraints, especially those defined by equalities, in machine-learning-based density functional design.
The content of this RSS Feed (c) The Royal Society of Chemistry

Digital Discovery, 2023, 2,1390-1403
DOI: 10.1039/D3DD00064H, Paper

Open Access

  This article is licensed under a Creative Commons Attribution 3.0 Unported Licence.

Jonathan P. Mailoa, Xin Li, Jiezhong Qiu, Shengyu Zhang
We use a mixture of incomplete-labelled molecule property databases to conditionally generate new molecules with multiple property co-constraints.
The content of this RSS Feed (c) The Royal Society of Chemistry

Digital Discovery, 2023, 2,1380-1389
DOI: 10.1039/D3DD00076A, Paper

Open Access

  This article is licensed under a Creative Commons Attribution-NonCommercial 3.0 Unported Licence.

Elvin Lo, Pin-Yu Chen
We benchmark zeroth-order methods for molecule optimization and discuss how they may be effectively used for practical molecule discovery.
The content of this RSS Feed (c) The Royal Society of Chemistry