Table 2.

Basic notation used in this work.

$\mathbb{R},\mathbb{N},\mathbb{B}$	The field for real, natural, and binary numbers
≃	It denotes isomorphism between two structures
$[N]$	The subset of natural numbers $\{1,\cdots ,N\}\subset \mathbb{N}$
$\mathcal{X},\mathbf{X},\mathbf{x},x$	Tensor, matrix, column vector, and scalar
$\dim (V)$	The dimension of a vector space, V
⊙	Hadamard product
⊗	Tensor product
${\times }_n$	The n-mode tensor-matrix product
${⊛}_N$	The Einstein product along the last N modes
$X^{(i_1,\cdots ,i_M)}$	The $(i_1,\cdots ,i_M)$th element of a sequence, $\{X^{(1,\cdots ,1)},\cdots ,X^{(I_1,\cdots ,I_M)}\}$
	indexed by $i_1\in [I_1],\cdots ,i_M\in [I_M]$, where X can be a scalar, vector, or tensor
${\mathcal{X}}_m$	The feature space for the m-th view
${\mathcal{Y}}_t$	The output space for the t-th task
$f_t$	The t-th classification task
$\mathit{g},\widehat{\mathit{g}}$	The MV-DTF layer and its low-rank approximation
$h_t$	The t-th task-specific function
${\mathcal{H}}_t$	Hypothesis space of the classifiers for the t-th task
$\mathcal{S}$	The latent tensor space
P	The order of the latent tensor space $\mathcal{S}$
$J_1\times \cdots \times J_P$	The dimension of the latent tensor space $\mathcal{S}$
$R^{(j_1,\cdots ,j_P)}$	For a tensor $\mathcal{A}\in {\mathbb{R}}^{J_1\times \cdots \times J_P\times I_1\times \cdots \times I_M}$, it denotes the
	tensor rank of the $(j_1,\cdots ,j_P)$-th subtensor ${\mathcal{A}}_{j_1\cdots j_P:\cdots :}$