scholarly journals Improving the accuracy of the neuroevolution machine learning potential for multi-component systems

2021 ◽  
Author(s):  
Zheyong Fan
Keyword(s):  
Neural Network ◽  
Machine Learning ◽  
Computational Cost ◽  
Md Simulations ◽  
Training Process ◽  
Learning Potential ◽  
Natural Evolution ◽  
Radial Functions ◽  
Component Systems ◽  
Training Neural Network

Abstract In a previous paper [Fan Z et al. 2021 Phys. Rev. B, 104, 104309], we developed the neuroevolution potential (NEP), a framework of training neural network based machine-learning potentials using a natural evolution strategy and performing molecular dynamics (MD) simulations using the trained potentials. The atom-environment descriptor in NEP was constructed based on a set of radial and angular functions. For multi-component systems, all the radial functions between two atoms are multiplied by some fixed factors that depend on the types of the two atoms only. In this paper, we introduce an improved descriptor for multi-component systems, in which different radial functions are multiplied by different factors that are also optimized during the training process, and show that it can significantly improve the regression accuracy without increasing the computational cost in MD simulations.

scholarly journals A machine-learning approach for identifying the counterparts of submillimetre galaxies and applications to the GOODS-North field

2019 ◽  
Vol 489 (2) ◽  
pp. 1770-1786 ◽  
Author(s):  
Ruihan Henry Liu ◽  
Ryley Hill ◽  
Douglas Scott ◽  
Omar Almaini ◽  
Fangxia An ◽  
...  
Keyword(s):  
Neural Network ◽  
Machine Learning ◽  
Near Infrared ◽  
Spectral Energy Distribution ◽  
Computational Cost ◽  
Spectral Energy ◽  
Distribution Fitting ◽  
Star Forming ◽  
Trained Neural Network ◽  
Validation Tests

ABSTRACT Identifying the counterparts of submillimetre (submm) galaxies (SMGs) in multiwavelength images is a critical step towards building accurate models of the evolution of strongly star-forming galaxies in the early Universe. However, obtaining a statistically significant sample of robust associations is very challenging due to the poor angular resolution of single-dish submm facilities. Recently, a large sample of single-dish-detected SMGs in the UKIDSS UDS field, a subset of the SCUBA-2 Cosmology Legacy Survey (S2CLS), was followed up with the Atacama Large Millimeter/submillimeter Array (ALMA), which has provided the resolution necessary for identification in optical and near-infrared images. We use this ALMA sample to develop a training set suitable for machine-learning (ML) algorithms to determine how to identify SMG counterparts in multiwavelength images, using a combination of magnitudes and other derived features. We test several ML algorithms and find that a deep neural network performs the best, accurately identifying 85 per cent of the ALMA-detected optical SMG counterparts in our cross-validation tests. When we carefully tune traditional colour-cut methods, we find that the improvement in using machine learning is modest (about 5 per cent), but importantly it comes at little additional computational cost. We apply our trained neural network to the GOODS-North field, which also has single-dish submm observations from the S2CLS and deep multiwavelength data but little high-resolution interferometric submm imaging, and we find that we are able to classify SMG counterparts for 36/67 of the single-dish submm sources. We discuss future improvements to our ML approach, including combining ML with spectral energy distribution fitting techniques and using longer wavelength data as additional features.

scholarly journals Gradient-Based Algorithm for Tracking the Activity of Neural Network Weights Changing

2021 ◽  
Vol 248 ◽  
pp. 01012
Author(s):  
Anton Starodub ◽  
Natalia Eliseeva ◽  
Milen Georgiev
Keyword(s):  
Neural Network ◽  
Machine Learning ◽  
Artificial Neural Network ◽  
Learning Process ◽  
Training Process ◽  
Important Indicator ◽  
Gradient Based ◽  
Artificial Neural ◽  
Training Efficiency ◽  
Learning Data

The research conducted in this paper is in the field of machine learning. The main object of the research is the learning process of an artificial neural network in order to increase its efficiency. The algorithm based on the analysis of retrospective learning data. The dynamics of changes in the values of the weights of an artificial neural network during training is an important indicator of training efficiency. The algorithm proposed in this work is based on changing the weight gradients values. Changing of the gradients weights makes it possible to understand how actively the network weights change during training. This knowledge helps to diagnose the training process and makes an adjusting the training parameters. The results of the algorithm can be used to train an artificial neural network. The network will help to determine the set of measures (actions) needed to optimize the learning process by the algorithm results.

scholarly journals Development and Application of a Single Neural Network Potential for IRMOF-n (n=1,4,6,7,10)

2021 ◽  
Author(s):  
Omer Tayfuroglu ◽  
Abdul Kadir Kocak ◽  
Yunus Zorlu
Keyword(s):  
Neural Network ◽  
Molecular Dynamics ◽  
Thermal Expansion ◽  
Density Of States ◽  
Mechanical Strain ◽  
Computational Cost ◽  
Negative Thermal Expansion ◽  
Md Simulations ◽  
Ab Initio Molecular Dynamics ◽  
Lattice Constants

Metal‑organic frameworks (MOFs) with their exceptional porous and organized structures have been subject of numerous applications. Predicting macroscopic properties from atomistic simulations require the most accurate force fields, which is still a major problem due to MOFs’ hybrid structures governed by covalent, ionic and dispersion forces. Application of ab‑initio molecular dynamics to such large periodic systems are thus beyond the current computational power. Therefore, alternative strategies must be developed to reduce computational cost without losing reliability. In this work, we describe the construction of a neural network potential (NNP) for IRMOF‑n series (n=1,4,7,10) trained by PBE-D4/def2-TZVP reference data of MOF fragments. We validated the resulting NNP on both fragments and bulk MOF structures by prediction of properties such as equilibrium lattice constants, phonon density of states and linker orientation. The energy and force RMSE values for the fragments are only 0.0017 eV/atom and 0.15 eV/Å, respectively. The NNP predicted equilibrium lattice constants of bulk structures, which are not included in training, are off by only 0.2-2.4% from experimental results. Moreover, our fragment trained NNP greatly predicts phenylene ring torsional energy barrier, equilibrium bond distances and vibrational density of states of bulk MOFs. Furthermore, NNP allows us to investigate unusual behaviors of selected MOFs such as the thermal expansion properties and the effect of mechanical strain on the adsorption of hydrogen and methane molecules. The NNP based molecular dynamics (MD) simulations suggest the IRMOF‑4 and IRMOF‑7 to have positive‑to‑negative thermal expansion coefficients while the rest to have only negative thermal expansion under the studied temperatures of 200 K to 400 K. The deformation of bulk structure by reduction of unit cell volume has shown to increase volumetric methane uptake in IRMOF‑1 but decrease in IRMOF‑7 due to the steric hindrance.

scholarly journals Implementation of Deep Learning for Slump Optimization Based on Concrete Quality Using Convolutional Neural Network in PT. SELO PROGO SAKTI

2020 ◽  
Vol 5 (2) ◽  
pp. 83-88
Author(s):  
Hedi Pandowo
Keyword(s):  
Neural Network ◽  
Machine Learning ◽  
Deep Learning ◽  
Experimental Design ◽  
Convolutional Neural Network ◽  
Processing Unit ◽  
Training Process ◽  
Average Accuracy ◽  
Graphical Processing ◽  
Intelligence Science

Deep Learning is part of the scientific field of Machine Learning and Machine Learning is part of Artificial Intelligence science. Deep Learning has extraordinary capabilities by using a hardware Graphical Processing Unit (GPU) so that the artificial requirement network can run faster than using a Personal Computer Unit (CPU). Especially in terms of object classification in images using existing methods in the Convolutional Neural Network (CNN). The method used in this research is Preprocessing and Processing of Input Data, Training Process in which CNN is trained to obtain high accuracy from the classification carried out and the Testing Process which is a classification process using weights and bias from the results of the training process. This type of research is a pre experimental design (pre experimental design). The results of the object image classification test with different levels of confusion in the Concrete database with the Mix Design K-125, K-150, K-250 and K-300 produce an average accuracy value. This is also relevant to measuring the failure rate of concrete or slump

COREFERENT PAIRS DETECTION IN UKRAINIAN TEXTS USING A CONVOLUTIONAL NEURAL NETWORK

2019 ◽  
Author(s):  
Sergiy Pogorilyy ◽  
Artem Kramov
Keyword(s):  
Neural Network ◽  
Machine Learning ◽  
Convolutional Neural Network ◽  
Language Processing ◽  
State Of The Art ◽  
Classification Problem ◽  
Machine Learning Algorithms ◽  
Coreference Resolution ◽  
Training Process ◽  
Natural Languages

The detection of coreferent pairs within a text is one of the basic tasks in the area of natural language processing (NLP). The state‑ of‑ the‑ art methods of coreference resolution are based on machine learning algorithms. The key idea of the methods is to detect certain regularities between the semantic or grammatical features of text entities. In the paper, the comparative analysis of current methods of coreference resolution in English and Ukrainian texts has been performed. The key disadvantage of many methods consists in the interpretation of coreference resolution as a classification problem. The result of coreferent pairs detection is the set of groups in which elements refer to a common entity. Therefore it is advisable to consider the coreference resolution as a clusterization task. The method of coreference resolution using the set of filtering sieves and a convolutional neural network has been suggested. The set of filtering sieves to find candidates for coreferent pairs formation has been implemented. The training process of a multichannel convolutional neural network on a marked Ukrainian corpus has been performed. The usage of a multichannel structure allows analyzing of the different components of text units: semantic, lexical, and grammatical features of words and sentences. Furthermore, it is possible to process input data with unfixed size (words or sentences of a text) using a convolutional layer. The output result of the method is the set of clusters. In order to form clusters, it is necessary to take into account the previous steps of the model’s workflow. Nevertheless, such an approach contradicts the traditional methodology of machine learning. Thus, the training process of the network has been performed using the SEARN algorithm that allows the solving of tasks with unfixed output structures using a classifier model. An experimental examination of the method on the corpus of Ukrainian news has been performed. In order to estimate the accuracy of the method the corresponding common metrics for clusterization tasks have been calculated. The results obtained can indicate that the suggested method can be used to find coreferent pairs within Ukrainian texts. The method can be also easily adapted and applied to other natural languages.

Application of Machine Learning in Turbulent Combustion for Aviation Gas Turbine Combustor Design

2021 ◽  
Author(s):  
Vishwas Verma ◽  
Kiran Manoharan ◽  
Jaydeep Basani
Keyword(s):  
Neural Network ◽  
Machine Learning ◽  
Flow Field ◽  
Gas Turbine ◽  
Computational Cost ◽  
Data Driven ◽  
Navier Stokes ◽  
Computational Time ◽  
Reacting Flow ◽  
Data Driven Approach

Abstract Numerical simulation of gas turbine combustors requires resolving a broad spectrum of length and time scales for accurate flow field and emission predictions. Reynold’s Averaged Navier Stokes (RANS) approach can generate solutions in few hours; however, it fails to produce accurate predictions for turbulent reacting flow field seen in general combustors. On the other hand, the Large Eddy Simulation (LES) approach can overcome this challenge, but it requires orders of magnitude higher computational cost. This limits designers to use the LES approach in combustor development cycles and prohibits them from using the same in numerical optimization. The current work tries to build an alternate approach using a data-driven method to generate fast and consistent results. In this work, deep learning (DL) dense neural network framework is used to improve the RANS solution accuracy using LES data as truth data. A supervised regression learning multilayer perceptron (MLP) neural network engine is developed. The machine learning (ML) engine developed in the present study can compute data with LES accuracy in 95% lesser computational time than performing LES simulations. The output of the ML engine shows good agreement with the trend of LES, which is entirely different from RANS, and to a reasonable extent, captures magnitudes of actual flow variables. However, it is recommended that the ML engine be trained using broad design space and physical laws along with a purely data-driven approach for better generalization.

scholarly journals A neural network protocol for electronic excitations of N-methylacetamide

2019 ◽  
pp. 201821044 ◽  
Author(s):  
Sheng Ye ◽  
Wei Hu ◽  
Xin Li ◽  
Jinxiao Zhang ◽  
Kai Zhong ◽  
...  
Keyword(s):  
Neural Network ◽  
Machine Learning ◽  
Ground State ◽  
Density Functional ◽  
Electronic Absorption Spectra ◽  
Dipole Moments ◽  
Computational Cost ◽  
Density Functional Theory Calculations ◽  
Peptide Bond ◽  
Transition Dipole

UV absorption is widely used for characterizing proteins structures. The mapping of UV spectra to atomic structure of proteins relies on expensive theoretical simulations, circumventing the heavy computational cost which involves repeated quantum-mechanical simulations of excited-state properties of many fluctuating protein geometries, which has been a long-time challenge. Here we show that a neural network machine-learning technique can predict electronic absorption spectra of N-methylacetamide (NMA), which is a widely used model system for the peptide bond. Using ground-state geometric parameters and charge information as descriptors, we employed a neural network to predict transition energies, ground-state, and transition dipole moments of many molecular-dynamics conformations at different temperatures, in agreement with time-dependent density-functional theory calculations. The neural network simulations are nearly 3,000× faster than comparable quantum calculations. Machine learning should provide a cost-effective tool for simulating optical properties of proteins.

scholarly journals A Study on the Effect of DropConnect to Control Overfitting in Designing Neural Networks

2020 ◽  
Author(s):  
Hyun-il Lim
Keyword(s):  
Neural Network ◽  
Machine Learning ◽  
Neural Networks ◽  
Neural Network Model ◽  
Network Models ◽  
Training Data ◽  
Neural Network Models ◽  
Structural Problems ◽  
The Neural Network ◽  
Training Neural Network

The neural network is an approach of machine learning by training the connected nodes of a model to predict the results of specific problems. The prediction model is trained by using previously collected training data. In training neural network models, overfitting problems can occur from the excessively dependent training of data and the structural problems of the models. In this paper, we analyze the effect of DropConnect for controlling overfitting in neural networks. It is analyzed according to the DropConnect rates and the number of nodes in designing neural networks. The analysis results of this study help to understand the effect of DropConnect in neural networks. To design an effective neural network model, the DropConnect can be applied with appropriate parameters from the understanding of the effect of the DropConnect in neural network models.

Exhaustive Probabilistic Neural Network for attribute selection and supervised seismic facies classification

2020 ◽  
pp. 1-67
Author(s):  
David Lubo-Robles ◽  
Thang Ha ◽  
Sivaramakrishnan Lakshmivarahan ◽  
Kurt J. Marfurt ◽  
Matthew J. Pranter
Keyword(s):  
Neural Network ◽  
Machine Learning ◽  
Probabilistic Neural Network ◽  
Learning Algorithms ◽  
Computational Cost ◽  
Component Analysis ◽  
Machine Learning Algorithms ◽  
Seismic Facies ◽  
Seismic Survey ◽  
Facies Classification

Machine learning algorithms such as principal component analysis (PCA), independent component analysis (ICA), self-organizing maps (SOM), and artificial neural networks (ANN), have been used by geoscientists to not only accelerate the interpretation of their data, but also to provide a more quantitative estimate of the likelihood that any voxel belongs to a given facies. Identifying the best combination of attributes needed to perform either supervised or unsupervised machine learning tasks continues to be the most-asked question by interpreters. In the past decades, stepwise regression and genetic algorithms have been used together with supervised learning algorithms to select the best number and combination of attributes. For reasons of computational efficiency, these techniques do not test all the seismic attribute combinations, potentially leading to a suboptimal classification. In this study, we develop an exhaustive probabilistic neural network (PNN) algorithm which exploits the PNN’s capacity in exploring non-linear relationships to obtain the optimal attribute subset that best differentiates target seismic facies of interest. We show the efficacy of our proposed workflow in differentiating salt from non-salt seismic facies in a Eugene Island seismic survey, offshore Louisiana. We find that from seven input candidate attributes, the Exhaustive PNN is capable of removing irrelevant attributes by selecting a smaller subset of four seismic attributes. The enhanced classification using fewer attributes also reduces the computational cost. We then use the resulting facies probability volumes to construct the 3D distribution of the salt diapir geobodies embedded in a stratigraphic matrix.

Sign in / Sign up

Export Citation Format

Share Document