Open Access
Table 2.
Architecture and hyperparameters of the BNN.
| Layer type | Activation function | |
|---|---|---|
| Dense, 1024 neurons | Leaky ReLu (α = 0.2) | |
| Dense, 1024 neurons | Leaky ReLu (α = 0.2) | |
| Dense, 1024 neurons | Leaky ReLu (α = 0.2) | |
| Dense Variational, 2 × (5 + 25) neurons | Linear | |
| Optimizer | Loss function | Batch size |
| RMSprop (learning rate = 5 × 10−4) | Log prob | 32 |
Notes. The neurons of the last layer predict the parameters of a mixture of two multivariate normal distributions.
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.