Thank you for visiting nature.com. You are using a browser version with limited support for CSS. To obtain
the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in
Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles
and JavaScript.
Opportunities for machine learning in chemical reaction networks
In this issue, Mingjian Wen, Kristin Persson and colleagues survey the different computational strategies available for chemical reaction network construction and analysis in a variety of applications, such as natural language processing and reaction property prediction. The opportunities for machine learning approaches, as well as the challenges that must still be overcome, are also discussed.
A proposed density functional approximation (DFA) recommender outperforms the use of a single functional by selecting the optimal exchange-correlation functional for a given system.
A framework for generating and interpreting dynamic visualizations from traditional static dimensionality reduction visualization methods has been proposed in a recent study.
Determining whether a drug candidate has sufficient affinity to its target is a critical part of drug development. A purely physics-based computational method was developed that uses non-equilibrium statistical mechanics approaches alongside molecular dynamics simulations. This technique could enable researchers to accurately estimate the binding affinities of potential drug candidates.
Chemical reaction networks are widely used to examine the behavior of chemical systems. While diverse strategies exist for chemical reaction network construction and analysis for a wide range of scientific goals, data-driven and machine learning methods must continue to capture increasingly complex phenomena to overcome existing unmet challenges.
Quantum algorithms for simulating quantum dynamics have shown promising results to overcome the difficulties from the classical counterparts. This Perspective summarizes the recent developments in the field, and further discusses the limitations and research opportunities towards the goal of quantum advantage.
A density functional recommender enables chemical space exploration by selecting the best exchange–correlation functional for each system, outperforming the use of a single functional for all systems or transfer learning models.
A hybrid functional (CF22D) with higher across-the-board accuracy for chemistry than most existing non-doubly hybrid functionals is presented by using a large database and a performance-triggered iterative supervised training method.
This work provides a physics-based theoretical framework for accurate protein–ligand binding affinity estimation based on molecular dynamics simulations, enhanced sampling, non-parametric reweighting and the orientation quaternion formalism.
Design choices for dimensionality reduction on calcium imaging recordings are systematically evaluated, and a method called calcium imaging linear dynamical system (CILDS) is proposed for performing deconvolution and dimensionality reduction jointly.
Data visualization is widely used in science, but interpreting such visualizations is prone to error. Here a dynamic visualization is introduced for capturing more information and improving the reliability of visual interpretations.
A computational workflow centered on probabilistic machine learning is developed to reconstruct the energy dispersion from photoemission band-mapping data. The workflow uncovers previously inaccessible momentum-space structural information at scale.