We present a set of techniques, based on the repeated arbitrary application of cp, cnot, and swap${}^{\ensuremath{\alpha}}$ (power-of-swap) quantum gate operations to an
n-qubit quantum computer that can be used in its verification. We find isomorphisms between the groups generated by these gate operations and known groups and use techniques from representation theory to determine their invariant subspaces. For the cp operation, we find an isomorphism to the direct product of $n(n\ensuremath{-}1)/2$ cyclic groups of order 2, and determine
2n one-dimensional invariant subspaces corresponding to the computational-basis vectors. For the cnot operation, we find an isomorphism to
GL(n,2), and determine two one-dimensional invariant subspaces and one $({2}^{n}\ensuremath{-}2)$-dimensional invariant subspace. For the swap${}^{\ensuremath{\alpha}}$ operation we find a complex structure of invariant subspaces with varying dimensions and occurrences and present a recursive procedure to construct them. Using knowledge of these invariant subspaces, we propose a hardware verification scheme which tests the correct functioning of a quantum computer.
We present a general condition to obtain subspaces that decay uniformly in a system governed by the Lindblad master equation and use them to perform error-mitigated quantum computation. The expectation values of dynamics encoded in such subspaces are unbiased estimators of noise-free expectation values. In analogy to the decoherence free subspaces which are left invariant by the action of Lindblad operators, we show that the uniformly decaying subspaces are left invariant (up to orthogonal terms) by the action of the dissipative part of the Lindblad equation. We apply our theory to a system of qubits and qudits undergoing relaxation with varying decay rates and show that such subspaces can be used to eliminate bias up to first-order variations in the decay rates without requiring full knowledge of noise. Since such a bias cannot be corrected through standard symmetry verification, our method can improve error mitigation in dual-rail qubits and, given partial knowledge of noise, can perform better than probabilistic error cancellation.