We consider a class of spin networks where each spin in a certain set interacts via Ising coupling with a single central spin. This is a common situation for instance in NV centers in diamonds. Due to the permutation symmetries of the network, the system is not globally controllable but it displays invariant subspaces of the underlying Hilbert space. The system is said to be subspace controllable if it is controllable on each of these subspaces. We characterize the given invariant subspaces and the dynamical Lie algebra of this class of systems and prove subspace controllability.
We consider the relation of symmetries and subspace controllability for spin-1/2 networks with XXZ couplings, subject to perturbation of a single node by a local potential (Z-control). The Hamiltonians for such networks decompose into excitation subspaces. Focusing on the single excitation subspace, it is shown for single-node Z-controls that external symmetries are characterized by eigenstates of the system Hamiltonian which have zero overlap with the control node, and there are no internal symmetries. It is further shown that there are symmetries which persist even in the presence of random perturbations. For XXZ chains with uniform coupling strengths, a characterization of all possible symmetries is given which shows a strong dependence on the position of the node we control. We then show for Heisenberg and XX chains with uniform coupling strength subject to single-node Z-control that the lack of symmetry is both necessary and sufficient for subspace controllability. Finally, the latter approach is generalized to establish controllability results for simple branched networks.
Abstract We present QNN-3L, a quantum neural network architecture based on three-level systems (qutrits) with longitudinal Ising-type couplings on a one-dimensional lattice. Starting from an explicit phenomenological multilevel Hamiltonian that includes local anharmonicity, diagonal couplings and resonant control pulses, we derive native selective rotations on computational subspaces and a nearest-neighbour controlled-
(Z) primitive. We prove that these native operations generate a subgroup dense in
(SU(3N)), providing approximate universality. To introduce effective nonlinearity while preserving coherent unitary propagation, we propose a hybrid architecture that uses a multi-subspace feature map: classical inputs are encoded via sequential rotations on two independent qutrit subspaces, whose non-commutativity enriches the implicit quantum kernel. We perform ablation studies and kernel diagnostics showing that the multi-subspace encoding yields improved kernel separation and more stable training than single-subspace alternatives. Numerical benchmarks for
(N=3)–6 qutrits (dimension
(dimH≤729)) demonstrate comparable expressibility at depths
(L≥3) and enhanced shallow-circuit entangling power relative to equivalent qubit circuits, and validate proof-of-concept supervised learning on a canonical nonlinear task. MSC Classification: 03.67.Ac , 03.67.Lx , 07.05.Mh