Table 1 Notations related to loss terms.

From: Deep hierarchical embedding for simultaneous modeling of GPCR proteins in a unified metric space

L Total loss to be used in training
\(L_{S}\) Summed cross-entropy loss with a softmax function from three hierarchies
\(L_{S_{i}}\) Cross-entropy loss with a softmax function from hierarchy level i where \(i \in \{cls, fam, sub\}\)
\(L_{C}\) Summed center loss from three hierarchies
\(L_{C_{i}}\) Center loss from hierarchy level i
\(d(x_i)\) Embedding vector in the hidden layer for input sequence \(x_i\)
\(\mu _{y_i}\) Class center of embedding vectors in deep feature space for class that input \(x_i\) belongs to