extended siamese neural networks (ESNN)

its a method to learn models based on two or more datapoints where one learns both an embedding function \(G(\vect{y}) = \hat{\vect{y}}\) and the contrastive function \(C(\vect{y},\vect{y})\)

\begin{equation} \label{orgd0c5efc} \mathbb{S}(\vect{x},\vect{y}) = C(G(\vect{x}),G(\vect{y})) , \end{equation}

This in contrast with siamese neural networks which only learns the unary embedding function \(G(\vect{y}) = \hat{\vect{y}}\).

See Mathisen2019

siamese neural networks

publications.bib