scholarly journals Convolutional Embedding of Attributed Molecular Graphs for Physical Property Prediction

2017 ◽  
Vol 57 (8) ◽  
pp. 1757-1772 ◽  
Author(s):  
Connor W. Coley ◽  
Regina Barzilay ◽  
William H. Green ◽  
Tommi S. Jaakkola ◽  
Klavs F. Jensen
Keyword(s):  
Physical Property ◽  
Property Prediction ◽  
Molecular Graphs

Physical Property Prediction for Waste Cooking Oil Biodiesel

2014 ◽  
Vol 7 (1) ◽  
pp. 62-68 ◽  
Author(s):  
Deng-Fang Ruan ◽  
Zhi-Hao Chen ◽  
Kui-Fang Wang ◽  
Yuan Chen ◽  
Fan Yang
Keyword(s):  
Surface Tension ◽  
Acid Methyl Ester ◽  
Physical Property ◽  
Waste Cooking Oil ◽  
Estimation Accuracy ◽  
Experimental Measurements ◽  
Property Prediction ◽  
Cooking Oil ◽  
Acid Methyl ◽  
Wide Range

This paper focuses on the physical property prediction of waste cooking oil biodiesel and examines the accuracy of different methods to estimate reliable basic physical properties, including density, viscosity and surface tension of waste cooking oil biodiesel in a wide range of temperature based on its fatty acid methyl ester component. A program for the physical property prediction of the biodiesel was developed and experimental measurements for the density, viscosity and surface tension of the biodiesel were performed to validate the chosen methods. The results show that the modified Rackett equation and the Orrick-Erbar method have the high estimation accuracy for the prediction of the density and the viscosity, respectively. The Sastri-Rao method and the Pizer method have the accuracy enough to predict the surface tension.

scholarly journals Path-Augmented Graph Transformer Network

2019 ◽  
Author(s):  
Benson Chen ◽  
Regina Barzilay ◽  
Tommi S Jaakkola
Keyword(s):  
Physical Chemistry ◽  
Quantum Chemistry ◽  
Recent Work ◽  
Higher Order ◽  
Molecular Property ◽  
Property Prediction ◽  
Molecular Graphs ◽  
Graph Properties ◽  
Local Aggregation ◽  
Aggregation Operations

<div>Much of the recent work on learning molecular representations has been based on Graph Convolution Networks (GCN). These models rely on local aggregation operations and can therefore miss higher-order graph properties. To remedy this, we propose Path-Augmented Graph Transformer Networks (PAGTN) that are explicitly built on longer-range dependencies in graphstructured data. Specifically, we use path features in molecular graphs to create global attention layers. We compare our PAGTN model against the GCN model and show that our model consistently</div><div>outperforms GCNs on molecular property prediction datasets including quantum chemistry (QM7, QM8, QM9), physical chemistry (ESOL, Lipophilictiy) and biochemistry (BACE, BBBP)2.</div>

scholarly journals Path-Augmented Graph Transformer Network

2019 ◽  
Author(s):  
Benson Chen ◽  
Regina Barzilay ◽  
Tommi S Jaakkola
Keyword(s):  
Physical Chemistry ◽  
Quantum Chemistry ◽  
Recent Work ◽  
Higher Order ◽  
Molecular Property ◽  
Property Prediction ◽  
Molecular Graphs ◽  
Graph Properties ◽  
Local Aggregation ◽  
Aggregation Operations

<div>Much of the recent work on learning molecular representations has been based on Graph Convolution Networks (GCN). These models rely on local aggregation operations and can therefore miss higher-order graph properties. To remedy this, we propose Path-Augmented Graph Transformer Networks (PAGTN) that are explicitly built on longer-range dependencies in graphstructured data. Specifically, we use path features in molecular graphs to create global attention layers. We compare our PAGTN model against the GCN model and show that our model consistently</div><div>outperforms GCNs on molecular property prediction datasets including quantum chemistry (QM7, QM8, QM9), physical chemistry (ESOL, Lipophilictiy) and biochemistry (BACE, BBBP)2.</div>

Correlations for Fluid Physical Property Prediction

1980 ◽  
Vol 32 (06) ◽  
pp. 968-970 ◽  
Author(s):  
M. Vasquez ◽  
H.D. Beggs
Keyword(s):  
Physical Property ◽  
Property Prediction
Sign in / Sign up

Export Citation Format

Share Document