Review and Prospect: Deep Learning in Nuclear Magnetic Resonance Spectroscopy

Since the concept of deep learning (DL) was formally proposed in 2006, it has had a major impact on academic research and industry. Nowadays, DL provides an unprecedented way to analyze and process data with demonstrated great results in computer vision, medical imaging, natural language processing, and so forth. Herein, applications of DL in NMR spectroscopy are summarized, and a perspective for DL as an entirely new approach that is likely to transform NMR spectroscopy into a much more efficient and powerful technique in chemistry and life sciences is outlined.     

Artificial Neural Networks for Pyrolysis,Thermal Analysis,and Thermokinetic Studies: The Status Quo

Artificial neural networks (ANNs) are a method of machine learning (ML) that is now widely used in physics, chemistry, and material science. ANN can learn from data to identify nonlinear trends and give accurate predictions. ML methods, and ANNs in particular, have already demonstrated their worth in solving various chemical engineering problems, but applications in pyrolysis, thermal analysis, and, especially, thermokinetic studies are still in an initiatory stage. The present article gives a critical overview and summary of the available literature on applying ANNs in the field of pyrolysis, thermal analysis, and thermokinetic studies. More than 100 papers from these research areas are surveyed. Some approaches from the broad field of chemical engineering are discussed as the venues for possible transfer to the field of pyrolysis and thermal analysis studies in general. It is stressed that the current thermokinetic applications of ANNs are yet to evolve significantly to reach the capabilities of the existing isoconversional and model-fitting methods.     

SERS-Based Biosensors Combined with Machine Learning for Medical Application**

Surface-enhanced Raman spectroscopy (SERS) has shown strength in non-invasive, rapid, trace analysis and has been used in many fields in medicine. Machine learning (ML) is an algorithm that can imitate human learning styles and structure existing content with the knowledge to effectively improve learning efficiency. Integrating SERS and ML can have a promising future in the medical field. In this review, we summarize the applications of SERS combined with ML in recent years, such as the recognition of biological molecules, rapid diagnosis of diseases, developing of new immunoassay techniques, and enhancing SERS capabilities in semi-quantitative measurements. Ultimately, the possible opportunities and challenges of combining SERS with ML are addressed.     

Artificial Intelligence for Autonomous Molecular Design: A Perspective

Domain-aware artificial intelligence has been increasingly adopted in recent years to expedite molecular design in various applications, including drug design and discovery. Recent advances in areas such as physics-informed machine learning and reasoning, software engineering, high-end hardware development, and computing infrastructures are providing opportunities to build scalable and explainable AI molecular discovery systems. This could improve a design hypothesis through feedback analysis, data integration that can provide a basis for the introduction of end-to-end automation for compound discovery and optimization, and enable more intelligent searches of chemical space. Several state-of-the-art ML architectures are predominantly and independently used for predicting the properties of small molecules, their high throughput synthesis, and screening, iteratively identifying and optimizing lead therapeutic candidates. However, such deep learning and ML approaches also raise considerable conceptual, technical, scalability, and end-to-end error quantification challenges, as well as skepticism about the current AI hype to build automated tools. To this end, synergistically and intelligently using these individual components along with robust quantum physics-based molecular representation and data generation tools in a closed-loop holds enormous promise for accelerated therapeutic design to critically analyze the opportunities and challenges for their more widespread application. This article aims to identify the most recent technology and breakthrough achieved by each of the components and discusses how such autonomous AI and ML workflows can be integrated to radically accelerate the protein target or disease model-based probe design that can be iteratively validated experimentally. Taken together, this could significantly reduce the timeline for end-to-end therapeutic discovery and optimization upon the arrival of any novel zoonotic transmission event. Our article serves as a guide for medicinal, computational chemistry and biology, analytical chemistry, and the ML community to practice autonomous molecular design in precision medicine and drug discovery.     

Structural, spectroscopic and magnetic properties of M[R2P(E)NP(E)R'2]2 complexes, M = Co, Mn, E = S, Se and R, R' = Ph or iPr. Covalency of M-S bonds from experimental data and theoretical calculations

The S/Se-containing bidentate ligands LH of the type R2P(E)NHP(E)R'2, E = S, Se and R, R' = Ph or iPr have been employed to synthesize ML2 (M = Mn, Co) complexes which contain the biologically important MS4 core. Theoretical calculations on the LH and L- forms of the ligands probe the geometric and electronic changes induced by the deprotonation of the LH form, which are correlated with structural data from X-ray crystallography. These results reflect the flexibility of the ligands, which enables them to be rather versatile with respect to the formation of ML2 complexes with varied geometries and MEPNPE metallacycle conformations. A series of old and new ML2 complexes have been synthesized and their structural, spectroscopic and magnetic properties characterized in detail. The nephelauxetic ratio beta of the CoL2 complexes provides evidence of covalent interactions, whereas the EPR properties of the MnL2 complexes are interpreted on the basis of predominant ionic interactions, between the metal center and the ligands, respectively. Additional evidence for the existence of covalent interactions in the CoL2 complexes (R = Ph, iPr, or mixed Ph/iPr), is offered by comparisons between their 31P NMR. The aforementioned notations are supported by extensive theoretical calculations on the ML2 (E = S, R = Me) modelled structures, which probe the covalent and ionic character of the M-S bonds when M = Co or Mn. Wider implications of the findings of the present study on the M-S covalency and its importance in the active sites of various metalloenzymes are also discussed.     

Hierarchically Structured Allotropes of Phosphorus from Data‐Driven Exploration

The discovery of materials is increasingly guided by quantum‐mechanical crystal‐structure prediction, but the structural complexity in bulk and nanoscale materials remains a bottleneck. Here we demonstrate how data‐driven approaches can vastly accelerate the search for complex structures, combining a machine‐learning (ML) model for the potential‐energy surface with efficient, fragment‐based searching. We use the characteristic building units observed in Hittorf's and fibrous phosphorus to seed stochastic (“random”) structure searches over hundreds of thousands of runs. Our study identifies a family of hierarchically structured allotropes based on a P8 cage as principal building unit, including one‐dimensional (1D) single and double helix structures, nanowires, and two‐dimensional (2D) phosphorene allotropes with square‐lattice and kagome topologies. These findings yield new insight into the intriguingly diverse structural chemistry of phosphorus, and they provide an example for how ML methods may, in the long run, be expected to accelerate the discovery of hierarchical nanostructures.     

Assessing conformer energies using electronic structure and machine learning methods

We have performed a large‐scale evaluation of current computational methods, including conventional small‐molecule force fields; semiempirical, density functional, ab initio electronic structure methods; and current machine learning (ML) techniques to evaluate relative single‐point energies. Using up to 10 local minima geometries across ~700 molecules, each optimized by B3LYP‐D3BJ with single‐point DLPNO‐CCSD(T) triple‐zeta energies, we consider over 6500 single points to compare the correlation between different methods for both relative energies and ordered rankings of minima. We find that the current ML methods have potential and recommend methods at each tier of the accuracy‐time tradeoff, particularly the recent GFN2 semiempirical method, the B97‐3c density functional approximation, and RI‐MP2 for accurate conformer energies. The ANI family of ML methods shows promise, particularly the ANI‐1ccx variant trained in part on coupled‐cluster energies. Multiple methods suggest continued improvements should be expected in both performance and accuracy.     

机器学习筛选用于气体吸附分离和存储的金属有机骨架材料

金属有机框架（Metal-organic framework ,MOF）因其高孔隙率、高比表面积和结构可调性,在气体吸附分离领域广泛应用。随着MOF数量激增,传统分子模拟和实验方法验证MOF性能成本高且速度慢,因此目前MOF筛选工作已转向高通量计算辅助的机器学习（Machine-learning,ML）。机器学习作为一种高效的大数据处理方法,能够在高通量筛选（High-Throughput Computational Screening,HTCS）的基础上对数据进行拟合,从而快速而准确地筛选出气体吸附分离材料,并深入挖掘其结构与性能之间的关系。本文回顾了近年机器学习应用于MOF筛选的研究。本文重点讨论了一些运用机器学习从大量结构中筛选出可用于CH4、H2和CO2等气体吸附分离与储存的MOF材料的工作。同时,我们梳理了当前MOF材料筛选工作中的研究思路和进展,并指出了机器学习在筛选MOF材料工作中面临的一些瓶颈和挑战。最后,对该领域的未来发展前景进行了展望。     

机器学习结合密度泛函理论计算在材料科学中的研究进展

近年来,材料科学研究中密度泛函理论(DFT)计算与机器学习相结合的方法呈现爆炸式增长的趋势。本文综述了DFT及其高通量方法产生的大量计算数据与机器学习相结合的原理和意义,从DFT计算的基本原理出发,重点介绍了机器学习方法的流程、常用的算法及其在催化材料预测热门研究方向中的应用,最后剖析了这个新兴领域目前存在的研究问题、挑战以及未来的发展前景。     

Benchmarking PES-Learn's machine learning models predicting accurate potential energy surface for quantum scattering

Machine learning (ML) models, neural networks, and Gaussian processes have been used to predict the potential energy surface taking C₂-He (both static and dynamic scenario) and NCCN-He collision systems. The surface is restricted to ∽125 points where traditional spline becomes inefficacious. Quantum dynamics is performed by solving close-coupling equation to compute cross sections benchmarking the performance of the ML models. The current study forms a basis for any future investigation of larger molecules where conventional fitting fails due to sparser ab initio points and cuts down the computational time without compromising on the quality of the surface.     

Synthesis, spectroscopic characterization and comparative DNA binding studies of Schiff base complexes derived from L-leucine and glyoxal

The mononuclear Schiff base complexes of the type, [ML(CH₃OH)₂] [M = Co(II), Ni(II), Cu(II) and Zn(II)] have been synthesized by template condensation of l-leucine and glyoxal. The complexes have been characterized on the basis of the results of the elemental analysis, molar conductance, magnetic susceptibility measurements and spectroscopic studies viz, FT-IR, Mass, ¹H NMR and ¹³C NMR spectra. The UV–vis and magnetic moment data revealed an octahedral geometry around Co(II), Ni(II) ion with distortion around Cu(II) ion complex confirmed by EPR data. The conductivity data show a non-electrolytic nature of the complexes. Absorption and fluorescence spectroscopic studies support that all the complexes exhibit a significant binding to calf thymus DNA.     

Synthesis and Characterization of Some New Triorganophosphorus, -Arsenic and -Antimony(V) Amido Derivatives

A number of triorganophosphorus, -arsenic, and -antimony(V) amido derivatives of the general formula R 2 R'ML 2 [where R = C 6 F 5 , C 6 H 5 ; R' = C 6 F 5 , C 6 H 4 CH 3 - p ; M = P, As or Sb and L = imidazole, benzimidazole, 2-methyl benzimidazole, indazole, and 1,2,3,4-tetrahydrocarbazole] have been synthesized by the metathetical reaction of triorganophosphorus, -arsenic, and -antimony(V) halides and the corresponding imidazoles in the presence of triethyl amine. These compounds have been characterized by elemental analysis, IR and NMR ( 1 H and 19 F) spectroscopy, conductance, and molecular weight data. The Van't Hoff factor "i" and molar conductance data of the compounds revealed them to be monomeric and nonionic in nature. On the basis of spectroscopic studies, a tentative trigonal bipyramidal structure has been assigned for these amido derivatives.     

Comparative study to predict toxic modes of action of phenols from molecular structures

Quantitative structure–activity relationship models for the prediction of mode of toxic action (MOA) of 221 phenols to the ciliated protozoan Tetrahymena pyriformis using atom-based quadratic indices are reported. The phenols represent a variety of MOAs including polar narcotics, weak acid respiratory uncouplers, pro-electrophiles and soft electrophiles. Linear discriminant analysis (LDA), and four machine learning techniques (ML), namely k-nearest neighbours (k-NN), support vector machine (SVM), classification trees (CTs) and artificial neural networks (ANNs), have been used to develop several models with higher accuracies and predictive capabilities for distinguishing between four MOAs. Most of them showed global accuracy of over 90%, and false alarm rate values were below 2.9% for the training set. Cross-validation, complementary subsets and external test set were performed, with good behaviour in all cases. Our models compare favourably with other previously published models, and in general the models obtained with ML techniques show better results than those developed with linear techniques. We developed unsupervised and supervised consensus, and these results were better than our ML models, the results of rule-based approach and other ensemble models previously published. This investigation highlights the merits of ML-based techniques as an alternative to other more traditional methods for modelling MOA.     

Science-Driven Atomistic Machine Learning

Machine learning (ML) algorithms are currently emerging as powerful tools in all areas of science. Conventionally, ML is understood as a fundamentally data-driven endeavour. Unfortunately, large well-curated databases are sparse in chemistry. In this contribution, I therefore review science-driven ML approaches which do not rely on “big data”, focusing on the atomistic modelling of materials and molecules. In this context, the term science-driven refers to approaches that begin with a scientific question and then ask what training data and model design choices are appropriate. As key features of science-driven ML, the automated and purpose-driven collection of data and the use of chemical and physical priors to achieve high data-efficiency are discussed. Furthermore, the importance of appropriate model evaluation and error estimation is emphasized.     

Synthesis and binding properties of calix[4]arenes with [2 + 2'] mixed ligating functional groups

A series of mixed [2 + 2'] p-tert-butylcalix[4]arene have been synthesised by selective 1,3-dialkylation of phenolic groups using various alkylating agents such as benzyl bromide, methyl iodide, ethyl bromoacetate, and 2-methoxyethyl tosylate. The extraction and complexation properties of the synthesized calixarenes towards alkali and alkaline earth metal cations have been investigated in acetonitrile by means of UV spectrophotometry and 1H NMR spectroscopy. The results show the formation of ML and/or ML2 species depending on the ligand and the cation. The enthalpies and entropies of complexation of alkali metal cations by a tetraglycol, diglycol-dibenzyl and diglycol-diester derivatives have been obtained from calorimetric measurements. The results revealed that the formation of ML species is controlled by enthalpy while the formation of ML2 from ML is entropy driven.     

人工智能在锂离子电池研发中的应用

锂离子电池已成为解决现代社会储能问题的最佳解决方案之一。然而,电池材料和器件开发都是复杂的多变量问题,传统的依赖研究人员进行实验的试错法在电池性能提升方面遇到了瓶颈。人工智能（AI）具有强大的高速、海量数据处理能力,是上述突破研究瓶颈的最具潜力的技术。其中,机器学习 (ML) 算法在评估多维数据变量和集合之间的组合关联方面的独特优势有望帮助研究人员发现不同因素之间的相互作用规律并阐明材料合成和设备制造的机制。本综述总结了锂离子电池传统研究方法遇到的各种挑战,并详细介绍了人工智能在电池材料研究、电池器件设计与制造、材料与器件表征、电池循环寿命与安全性评估等方面的应用。最重要的是,我们介绍了AI和ML在电池研究中面临的挑战,并讨论了它们应用的缺点和前景。我们相信,未来实验科学家、数学建模专家和AI专家之间更紧密的合作将极大地促进AI和ML方法用以解决传统方法难以克服的电池和材料问题。     

Accelerated Nuclear Magnetic Resonance Spectroscopy with Deep Learning

Nuclear magnetic resonance (NMR) spectroscopy serves as an indispensable tool in chemistry and biology but often suffers from long experimental times. We present a proof-of-concept of the application of deep learning and neural networks for high-quality, reliable, and very fast NMR spectra reconstruction from limited experimental data. We show that the neural network training can be achieved using solely synthetic NMR signals, which lifts the prohibiting demand for a large volume of realistic training data usually required for a deep learning approach.     

Accelerated Nuclear Magnetic Resonance Spectroscopy with Deep Learning

Nuclear magnetic resonance (NMR) spectroscopy serves as an indispensable tool in chemistry and biology but often suffers from long experimental times. We present a proof‐of‐concept of the application of deep learning and neural networks for high‐quality, reliable, and very fast NMR spectra reconstruction from limited experimental data. We show that the neural network training can be achieved using solely synthetic NMR signals, which lifts the prohibiting demand for a large volume of realistic training data usually required for a deep learning approach.