표·그림DB
Fig. 3.31. Manifold of the embedding vector spaces colorized by three human-annotated label data sets: temporal sequence (top), concept evaluation label (middle), and intensity level of predicted Rainfall class (bottom). The embedding vectors from multi-label soft-silhouette deep clustering (z̃soft) exclusively represent temporally and conceptually disentangled features. However, the given embedding vectors (z) are only highly dependent on the predictive rainfall intensity classes, which are related to its objective function.
표·그림명
Fig. 3.31. Manifold of the embedding vector spaces colorized by three human-annotated label data sets: temporal sequence (top), concept evaluation label (middle), and intensity level of predicted Rainfall class (bottom). The embedding vectors from multi-label soft-silhouette deep clustering (z̃soft) exclusively represent temporally and conceptually disentangled features. However, the given embedding vectors (z) are only highly dependent on the predictive rainfall intensity classes, which are related to its objective function.