Table 4
Overview of learning paradigms and architectures and their corresponding MAE for various labels.
| (a) Model descriptions | |||||
|---|---|---|---|---|---|
| Model type | Learning paradigm | Architecture | Bottleneck | Dataset | |
| Legend descriptor | |||||
| β-VAE (U*) (1) | β-VAE | Unsupervised | CNN–CNN | 128 | Real |
| CNN (mix) | Encoder | Supervised | CNN | 6 | Mixed |
| AE (real) | AE | Semi-supervised | CNN-ResNet | 9 | Real |
| AE (mix) | AE | Semi-supervised | CNN-ResNet | 9 | Mixed |
| infoVAE (mix) | infoVAE | Semi-supervised | CNN-ResNet | 32 | Mixed |
| CNN (ETC) | Encoder | Supervised | CNN | 7 | ETC |
| (b) Intrinsic labels | |||||
| Teff (K) | [M/H] (dex) | log(g) (dex) | All (−) | ||
| β-VAE (U*) (1) | 978.0 ± 13.0 | 0.1578 ± 0.0027 | 0.1798 ± 0.0025 | 1.548 ± 0.013 | |
| CNN (mix) | 50.21 ± 0.82 | 0.02118 ± 0.00064 | 0.03957 ± 0.00085 | 0.1793 ± 0.0022 | |
| AE (real) | 50.32 ± 0.83 | 0.02478 ± 0.00056 | 0.03898 ± 0.00089 | 0.1844 ± 0.0022 | |
| AE (mix) | 49.74 ± 0.77 | 0.02292 ± 0.00055 | 0.04139 ± 0.00087 | 0.1867 ± 0.0021 | |
| infoVAE (mix) | 43.68 ± 0.73 | 0.03424 ± 0.00058 | 0.04005 ± 0.00087 | 0.1986 ± 0.0021 | |
| CNN (ETC) | 273.1 ± 4.2 | 0.1222 ± 0.0011 | 0.4994 ± 0.0044 | 1.7029 ± 0.0091 | |
| (c) Extrinsic labels | |||||
| Radvel (km s−1) | BERV (km s−1) | Airmass (−) | All (−) | ||
| β-VAE (U*) (1) | 31.07 ± 0.27 | 16.01 ± 0.096 | 0.181 ± 0.0019 | 1.5751 ± 0.0074 | |
| CNN (mix) | 2.375 ± 0.061 | 0.1671 ± 0.002 | 0.01069 ± 0.00022 | 0.0815 ± 0.001 | |
| AE (real) | 1.94 ± 0.069 | 0.1214 ± 0.0021 | 0.00885 ± 0.00018 | 0.0664 ± 0.0011 | |
| AE (mix) | 1.918 ± 0.063 | 0.1697 ± 0.0024 | 0.0103 ± 0.00022 | 0.0709 ± 0.0011 | |
| infoVAE (mix) | 1.967 ± 0.071 | 0.1855 ± 0.0024 | 0.01023 ± 0.00021 | 0.0722 ± 0.0011 | |
| CNN (ETC) | 25.43 ± 0.27 | NaN ± NaN | 1.6827 ± 0.0054 | 6.932 ± 0.016 | |
Notes. The column titled “All” uses NMAE from Eq. (13) to summarize all labels in the group. The notation “a ± b” represents the mean absolute error ± the standard deviation of the estimate for each respective label and model. Using boldface signifies that the model outperformed the other models significantly for the label in that column. This significance was determined with a significance level of α = 0.01 using Mann-Whitney U tests with Holm–Bonferroni correction for multiple comparisons (Mann & Whitney 1947; Holm 1979). (1) Sedaghat et al. (2021).
Current usage metrics show cumulative count of Article Views (full-text article views including HTML views, PDF and ePub downloads, according to the available data) and Abstracts Views on Vision4Press platform.
Data correspond to usage on the plateform after 2015. The current usage metrics is available 48-96 hours after online publication and is updated daily on week days.
Initial download of the metrics may take a while.