Table 1 MAE Errors in the reflection and transmission without no simulations included in the training data and 5 simulations included in training.

From: Deep learning based analysis of microstructured materials for thermal radiation control

Material

MAE reflection

MAE transmission

Error difference

No simulations

5 Simulations

No simulations

5 Simulations

ΔR

ΔT

Au88

0.0637

0.0456

2.63E−05

3.88E−05

− 0.018

1.25E−05

B4C89

0.0394

0.0098

0.0542

0.0091

− 0.030

− 0.045

BaF290

0.0040

0.0009

0.0907

0.0053

− 0.003

− 0.085

Be91

0.0086

0.0067

5.01E−06

3.72E−06

− 0.002

− 1.29E−06

C92

0.0972

0.0124

3.11E−04

3.12E−05

− 0.085

− 2.80E−04

Cu79

0.0709

0.0880

1.07E−02

1.80E−05

0.017

− 0.011

GaAs79

0.0249

0.0106

0.0927

0.0186

− 0.014

− 0.074

Ge79

0.0869

0.0216

0.0964

0.0094

− 0.065

− 0.087

In93

0.0312

0.0223

6.95E−06

4.30E−06

− 0.009

− 2.65E−06

InAs94

0.0036

0.0030

0.0849

0.0121

− 0.001

− 0.073

InP94

0.0098

0.0053

0.0722

0.0126

− 0.005

− 0.060

Li95

0.0570

0.0297

8.23E−04

8.21E−04

− 0.027

− 2.04E−06

Mg96

0.0434

0.0295

5.84E−06

4.14E−06

− 0.014

− 1.69E−06

Mo82

0.0316

0.0296

7.39E−06

5.04E−06

− 0.002

− 2.35E−06

Nb97

0.0156

0.0154

4.55E−06

3.13E−06

0.000

− 1.42E−06

Os98

0.0282

0.0291

1.21E−04

1.20E−04

0.001

− 1.70E−06

Pd91

0.0168

0.0109

1.48E−05

1.02E−05

− 0.006

− 4.56E−06

Pt79

0.0096

0.0086

5.49E−06

4.32E−06

− 0.001

− 1.17E−06

Rh79

0.0106

0.0125

8.07E−06

6.89E−06

0.002

− 1.18E−06

Si3N487

0.0021

0.0021

0.0321

0.0046

0.000

− 0.028

TiO2

0.0006

0.0016

0.2003

0.0359

0.001

− 0.164

Zn90

0.0554

0.0203

1.77E−05

1.34E−05

− 0.035

− 4.31E−06

Zr90

0.0141

0.0122

3.21E−05

2.17E−05

− 0.002

− 1.04E−05

  1. The relative error for each material is shown, with nearly all materials showing a significant decrease in error as a result of the small amount of data being included.