Table 4 Measured MAEs of the competitor methods and NN-CroMEL on the target experimental datasets

From: Cross-modality material embedding loss for transferring knowledge between heterogeneous material descriptors

Target Experimental Dataset

XGBoost

FDNN

ElemNet

RFN

Roost

NN-CroMEL

ESTM-RT-PF

0.536 (0.081)

0.726 (0.037)

1.266 (0.144)

1.267 (0.160)

0.640 (0.099)

0.506 (0.105)

ESTM-RT-ZT

0.044 (0.005)

0.053 (0.007)

0.066 (0.010)

0.059 (0.009)

0.055 (0.009)

0.035 (0.008)

ESTM-HT-PF

0.329 (0.104)

0.424 (0.080)

1.252 (0.133)

1.367 (0.126)

0.383 (0.056)

0.238 (0.054)

ESTM-HT-ZT

0.100 (0.013)

0.113 (0.014)

0.144 (0.014)

0.141 (0.016)

0.134 (0.021)

0.085 (0.012)

SCMAT-BIN

1.432 (0.044)

1.486 (0.037)

1.579 (0.050)

1.519 (0.083)

1.389 (0.027)

0.992 (0.190)

SCMAT-TER

1.951 (0.086)

2.168 (0.113)

3.105 (0.207)

2.681 (0.100)

2.222 (0.106)

1.418 (0.272)

IPOP-200-QE

8.343 (1.780)

9.495 (0.569)

13.241 (2.457)

13.625 (3.133)

10.132 (1.087)

7.367 (0.957)

IPOP-200-DT

0.630 (0.244)

0.819 (0.147)

2.119 (0.412)

1.822 (0.327)

0.583 (0.202)

0.493 (0.193)

IPOP-300-QE

8.885 (0.518)

13.480 (0.644)

12.749 (0.899)

13.927 (1.342)

12.373 (0.903)

9.229 (1.883)

IPOP-300-DT

0.606 (0.100)

1.044 (0.159)

2.001 (0.100)

1.412 (0.083)

0.788 (0.118)

0.548 (0.122)

Liverpool-RT

2.136 (0.296)

2.391 (0.230)

2.720 (0.354)

2.815 (0.308)

2.535 (0.432)

1.710 (0.361)

MGTT

0.020 (0.003)

0.022 (0.002)

0.041 (0.013)

0.022 (0.002)

0.024 (0.003)

0.015 (0.001)

EFE

0.074 (0.003)

0.064 (0.004)

0.095 (0.003)

0.094 (0.003)

0.065 (0.002)

0.038 (0.013)

EBG

0.309 (0.012)

0.326 (0.008)

0.466 (0.021)

0.439 (0.005)

0.312 (0.013)

0.182 (0.056)

  1. The standard deviations of the measured MAEs are presented in the parentheses, and the smallest MAE for each target experimental dataset is highlighted by bold font.