We provide a generalization of the idea of unitary designs to cover finite averaging over much more general operations on quantum states. Namely, we construct finite averaging sets for averaging quantum states over arbitrary reductive Lie groups, on condition that the averaging is performed uniformly over the compact component of the group. Our construction comprises probabilistic mixtures of unitary 1-designs on specific operator subspaces. Provided construction is very general, competitive in the size of the averaging set when compared to other known constructions, and can be efficiently implemented in the quantum circuit model of computation.
We provide a generalization of the idea of unitary designs to cover finite averaging over much more general operations on quantum states. Namely, we construct finite averaging sets for averaging quantum states over arbitrary reductive Lie groups, on condition that the averaging is performed uniformly over the compact component of the group. Our construction comprises probabilistic mixtures of unitary 1-designs on specific operator subspaces. Provided construction is very general, competitive in the size of the averaging set when compared to other known constructions, and can be efficiently implemented in the quantum circuit model of computation.
We construct finite averaging sets for twirling finite-dimensional quantum states over a broad class of symmetry transformations represented by reductive Lie groups. The averaging is defined through the Cartan decomposition of the group, with Haar-uniform integration over the compact components and an arbitrary normalized measure over the non-compact Abelian component. We show that the resulting twirling map can always be represented as a finite probabilistic mixture of unitary 1-design channels acting on irreducible sectors of the representation restricted to the maximal compact subgroup. Thus, the finite construction is determined by the compact-sector decomposition, while the non-compact Cartan component contributes only scalar sector weights. This distinction is important for implementation: no Schur transform for the full reductive group is required. Compared with standard t -design-based finite averaging, our construction generally uses fewer averaging operators, at the price of replacing local tensor-power sampling by sector-wise unitary averaging. This could be especially beneficial for simulations and quantum algorithms on higher-dimensional systems as reduction of the size of the averaging set relatively increases with dimension.
Abstract Averaging physical quantities over Lie groups appears in many contexts across the rapidly developing branches of physics like quantum information science or quantum optics. Such an averaging process can be always represented as averaging with respect to a finite number of elements of the group, called a finite averaging set . In the previous research such sets, known as t -designs, were constructed only for the case of averaging over unitary groups (hence the name unitary t -designs ). In this work we investigate the problem of constructing finite averaging sets for averaging over general non-compact matrix Lie groups, which is much more subtle task due to the fact that the the uniform invariant measure on the group manifold (the Haar measure) is infinite. We provide a general construction of such sets based on the Cartan decomposition of the group, which splits the group into its compact and non-compact components. The averaging over the compact part can be done in a uniform way, whereas the averaging over the non-compact one has to be endowed with a suppressing weight function, and can be approached using generalized Gauss quadratures. This leads us to the general form of finite averaging sets for semisimple matrix Lie groups in the product form of finite averaging sets with respect to the compact and non-compact parts. We provide an explicit calculation of such sets for the group S L ( 2 , C ) , although our construction can be applied to other cases. Possible applications of our results cover finding finite ensembles of random operations in quantum information science and quantum optics, which can be used in constructions of randomised quantum algorithms, including optical interferometric implementations.
Abstract It is a well-established fact in quantum information theory, that uniform averaging over the collective action of a unitary group on a multipartite quantum state projects the state to a form equivalent to a permutation operator of the subsystems. Hence states equivalent to permutation operators are untouched by collective unitary noise. A trivial observation shows that uniform averaging over permutation operators projects the state into a form with block-diagonal structure equivalent to the one of the collective action of the unitary group. We introduce a name for this property: duality of averaging . The mathematical reason behind this duality is the fact that the collective action of the unitary group on the tensor product state space of a multipartite quantum system and the action of the permutation operations are mutual commutants when treated as matrix algebras. Such pairs of matrix algebras are known as dual reductive pairs . In this work we show, that in the case of finite dimensional quantum systems such duality of averaging holds for any pairs of symmetry groups being dual reductive pairs, regardless of whether they are compact or not, as long as the averaging operation is defined via iterated integral over the Cartan decomposition of the group action. Although our result is very general, we focus much attention on the concrete example of a dual reductive pair consisting of collective action of special linear matrices and permutation operations, which physically corresponds to averaging multipartite quantum states over non-unitary SLOCC-type (Stochastic Local Operations and Classical Communication) operations. In this context we show, that noiseless subsystems known from collective unitary averaging persist in the case of SLOCC averaging in a conditional way: upon postselection to specific invariant subspaces.
Unitary designs are essential tools in several quantum information protocols. Similarly to other design concepts, unitary designs are mainly used to facilitate averaging over a relevant space, in this case, the unitary group U ( d ) . While it is known that exact unitary t -designs exist for any degree t and dimension d , the most appealing type of designs, group designs (in which the elements of the design form a group), can provide at most 3 -designs. Moreover, even group 2 -designs can exist only in limited dimensions. In this paper, we present novel construction methods for creating exact generalized group designs based on the representation theory of the unitary group and its finite subgroups that overcome the 4 -design-barrier of unitary group designs. Furthermore, a construction is presented for creating generalized group 2 -designs in arbitrary dimensions.
We investigate the generation of quantum states and unitary operations that are ``random'' in certain respects. We show how to use such states to estimate the average fidelity, an important measure in the study of implementations of quantum algorithms. We re-discover the result that the states of a maximal set of mutually-unbiased bases serve this purpose. An efficient circuit is presented that generates an arbitrary state out of such a set. Later on, we consider unitary operations that can be used to turn any quantum channel into a depolarizing channel. It was known before that the Clifford group serves this and a related purpose, and we show that these are actually the same. We also show that a small subset of the Clifford group is already sufficient to accomplish this. We conclude with an efficient construction of the elements of that subset. This thesis is based on joint work with Richard Cleve, Joseph Emerson, and Etera Livine.
LetUdbe a unitary operator representing an arbitraryd-dimensional unitary quantum operation. This work presents optimal quantum circuits for transforming a numberkof calls ofUdinto its complex conjugateUd. Our circuits admit a parallel implementation and are proven to be optimal for anykanddwith an average fidelity of⟨F⟩=d(d−k)k+1. Optimality is shown for average fidelity, robustness to noise, and other standard figures of merit. This extends previous works which considered the scenario of a single call (k=1) of the operationUd, and the special case ofk=d−1calls. We then show that our results encompass optimal transformations fromkcalls ofUdtof(Ud)for any arbitrary homomorphismffrom the group ofd-dimensional unitary operators to itself, since complex conjugation is the only non-trivial automorphism on the group of unitary operators. Finally, we apply our optimal complex conjugation implementation to design a probabilistic circuit for reversing arbitrary quantum evolutions.