2019
DOI: 10.1002/advs.201801367
|View full text |Cite
|
Sign up to set email alerts
|

Deep Learning Spectroscopy: Neural Networks for Molecular Excitation Spectra

Abstract: Deep learning methods for the prediction of molecular excitation spectra are presented. For the example of the electronic density of states of 132k organic molecules, three different neural network architectures: multilayer perceptron (MLP), convolutional neural network (CNN), and deep tensor neural network (DTNN) are trained and assessed. The inputs for the neural networks are the coordinates and charges of the constituent atoms of each molecule. Already, the MLP is able to learn spectra, but the root mean sq… Show more

Help me understand this report

Search citation statements

Order By: Relevance

Paper Sections

Select...
4
1

Citation Types

1
244
0
1

Year Published

2019
2019
2024
2024

Publication Types

Select...
7

Relationship

1
6

Authors

Journals

citations
Cited by 231 publications
(246 citation statements)
references
References 66 publications
1
244
0
1
Order By: Relevance
“…[50,[247][248][249][250][251][252][253][254] Apart from such generic approaches and packages, machine learning models are often customized to materials science. One example is the creation of custom neural network architectures [201,204,206,255,256] that have been designed specifically for atomistic geometries, reducing the need for feature engineering.…”
Section: Specifics Of Machine Learning In Materials Sciencementioning
confidence: 99%
See 1 more Smart Citation
“…[50,[247][248][249][250][251][252][253][254] Apart from such generic approaches and packages, machine learning models are often customized to materials science. One example is the creation of custom neural network architectures [201,204,206,255,256] that have been designed specifically for atomistic geometries, reducing the need for feature engineering.…”
Section: Specifics Of Machine Learning In Materials Sciencementioning
confidence: 99%
“…[17] Supervised learning applies in situations where a machine learning model is trained on input-output pairs from a real process to produce optimal outputs for unseen inputs. Typical applications are predictions of physical properties (like formation energies [200][201][202] or molecular properties [203][204][205][206][207] ) given the input features of a material or process (e.g., geometry, physical properties, external conditions).…”
Section: Introduction To Machine Learningmentioning
confidence: 99%
“…In contrast to an artificial neural network, the convolution layer in CNN allows the model to extract small details and to be trained on the extracted details of the input data, which improves its prediction accuracy . Since then, CNNs have been widely used for image recognition or image classification, and there were several trials to use artificial intelligence algorithm to study Raman spectral data as well as different types of spectroscopic data . In this study, we suggest a platform for Raman signature classification of EVs based on CNN.…”
Section: Introductionmentioning
confidence: 99%
“…[36,37] Since then, CNNs have been widely used for image recognition or image classification, and there were several trials to use artificial intelligence algorithm to study Raman spectral data [27][28][29][30][31]38] as well as different types of spectroscopic data. [39][40][41][42] In this study, we suggest a platform for Raman signature classification of EVs based on CNN. The classification performed in this article is aimed at finding the spectral differences between prostate cancer derived EVs and blood cell derived EVs, because the latter are the clinically relevant background of the measurement.…”
Section: Introductionmentioning
confidence: 99%
“…Deep neural networks (DNNs) are currently state-of-the-art in image recognition applications, and have already been tested for several scientific spectroscopy applications [1][2][3][4]. In fact, fast machine learning processing will be crucial for high-throughput data analysis, especially for large research experiment facilities such as synchrotron or free-electron lasers [5], where the large data amount prevents the standard hand processing.…”
Section: Introductionmentioning
confidence: 99%