Explore our research publications
Explore our research publications
Explore our research publications
Pioneering
Quantum Insights
Pioneering
Quantum Insights
Discover a curated selection of publications showcasing our breakthroughs in quantum research and computational science. Stay updated with fresh insights and innovations from our expert team.
Discover a curated selection of publications showcasing our breakthroughs in quantum research and computational science. Stay updated with fresh insights and innovations from our expert team.
Discover a curated selection of publications showcasing our breakthroughs in quantum research and computational science. Stay updated with fresh insights and innovations from our expert team.
Novel oracle constructions for quantum random access memory
Ákos Nagy, Cindy Zhang
A New Approach to Quantum Random Access Memory (QRAM) Discover our innovative design for QRAM that leverages the Walsh-Hadamard Transform to optimize circuit efficiency. Our method adapts the complexity of the circuit based on the transform's sparsity rather than the function itself—ideal for binary optimization and other low-degree functions. With a flexible design that lets you trade circuit depth for qubit count, our QRAM circuits achieve significant performance improvements, even for challenging Boolean functions. Download the full paper to learn more about these scalable and efficient QRAM constructions.
Download
Novel oracle constructions for quantum random access memory
Ákos Nagy, Cindy Zhang
A New Approach to Quantum Random Access Memory (QRAM) Discover our innovative design for QRAM that leverages the Walsh-Hadamard Transform to optimize circuit efficiency. Our method adapts the complexity of the circuit based on the transform's sparsity rather than the function itself—ideal for binary optimization and other low-degree functions. With a flexible design that lets you trade circuit depth for qubit count, our QRAM circuits achieve significant performance improvements, even for challenging Boolean functions. Download the full paper to learn more about these scalable and efficient QRAM constructions.
Download
Novel oracle constructions for quantum random access memory
Ákos Nagy, Cindy Zhang
A New Approach to Quantum Random Access Memory (QRAM) Discover our innovative design for QRAM that leverages the Walsh-Hadamard Transform to optimize circuit efficiency. Our method adapts the complexity of the circuit based on the transform's sparsity rather than the function itself—ideal for binary optimization and other low-degree functions. With a flexible design that lets you trade circuit depth for qubit count, our QRAM circuits achieve significant performance improvements, even for challenging Boolean functions. Download the full paper to learn more about these scalable and efficient QRAM constructions.
Download
Diagonal operator decomposition on restricted topologies via enumeration of quantum state subsets
Jan Tułowiecki, Łukasz Czerwiński, Konrad Deka, Jan Gwinner, Witold Jarnicki, Adam Szady
Various quantum algorithms require usage of arbitrary diagonal operators as subroutines. For their execution on a physical hardware, those operators must be first decomposed into target device's native gateset and its qubit connectivity for entangling gates. Here, we assume that the allowed gates are exactly the CX gate and the parameterized phase gate. We introduce a framework for the analysis of CX-only circuits and through its lens provide solution constructions for several different device topologies (fully-connected, linear and circular). We also introduce two additional variants of the problem. Those variants can be used in place of exact decomposition of the diagonal operator when the circuit following it satisfies a set of prerequisites, enabling further reduction in the CX cost of implementation. Finally, we discuss how to exploit the framework for the decomposition of a particular, rather than general, diagonal operator.
Download
Diagonal operator decomposition on restricted topologies via enumeration of quantum state subsets
Jan Tułowiecki, Łukasz Czerwiński, Konrad Deka, Jan Gwinner, Witold Jarnicki, Adam Szady
Various quantum algorithms require usage of arbitrary diagonal operators as subroutines. For their execution on a physical hardware, those operators must be first decomposed into target device's native gateset and its qubit connectivity for entangling gates. Here, we assume that the allowed gates are exactly the CX gate and the parameterized phase gate. We introduce a framework for the analysis of CX-only circuits and through its lens provide solution constructions for several different device topologies (fully-connected, linear and circular). We also introduce two additional variants of the problem. Those variants can be used in place of exact decomposition of the diagonal operator when the circuit following it satisfies a set of prerequisites, enabling further reduction in the CX cost of implementation. Finally, we discuss how to exploit the framework for the decomposition of a particular, rather than general, diagonal operator.
Download
Diagonal operator decomposition on restricted topologies via enumeration of quantum state subsets
Jan Tułowiecki, Łukasz Czerwiński, Konrad Deka, Jan Gwinner, Witold Jarnicki, Adam Szady
Various quantum algorithms require usage of arbitrary diagonal operators as subroutines. For their execution on a physical hardware, those operators must be first decomposed into target device's native gateset and its qubit connectivity for entangling gates. Here, we assume that the allowed gates are exactly the CX gate and the parameterized phase gate. We introduce a framework for the analysis of CX-only circuits and through its lens provide solution constructions for several different device topologies (fully-connected, linear and circular). We also introduce two additional variants of the problem. Those variants can be used in place of exact decomposition of the diagonal operator when the circuit following it satisfies a set of prerequisites, enabling further reduction in the CX cost of implementation. Finally, we discuss how to exploit the framework for the decomposition of a particular, rather than general, diagonal operator.
Download
Clifford circuits over non-cyclic abelian groups
Milo Moses, Jacek Horecki, Konrad Deka, Jan Tulowiecki
We present a discussion of the generalized Clifford group over non-cyclic finite abelian groups. These Clifford groups appear naturally in the theory of topological error correction and abelian anyon models. We demonstrate a generalized Gottesman-Knill theorem, stating that every Clifford circuit can be efficiently classically simulated. We additionally provide circuits for a universal quantum computing scheme based on local two-qudit Clifford gates and magic states.
Download
Clifford circuits over non-cyclic abelian groups
Milo Moses, Jacek Horecki, Konrad Deka, Jan Tulowiecki
We present a discussion of the generalized Clifford group over non-cyclic finite abelian groups. These Clifford groups appear naturally in the theory of topological error correction and abelian anyon models. We demonstrate a generalized Gottesman-Knill theorem, stating that every Clifford circuit can be efficiently classically simulated. We additionally provide circuits for a universal quantum computing scheme based on local two-qudit Clifford gates and magic states.
Download
Clifford circuits over non-cyclic abelian groups
Milo Moses, Jacek Horecki, Konrad Deka, Jan Tulowiecki
We present a discussion of the generalized Clifford group over non-cyclic finite abelian groups. These Clifford groups appear naturally in the theory of topological error correction and abelian anyon models. We demonstrate a generalized Gottesman-Knill theorem, stating that every Clifford circuit can be efficiently classically simulated. We additionally provide circuits for a universal quantum computing scheme based on local two-qudit Clifford gates and magic states.
Download
Fixed-point Grover Adaptive Search for Quadratic Binary Optimization Problems
Ákos Nagy, Jaime Park, Cindy Zhang, Atithi Acharya, Alex Khan
We explore a variant of Grover’s search tailored for solving Quadratic Unconstrained Binary Optimization (QUBO) problems. For a QUBO problem with n dimensions and m nonzero terms, we design a “marker” oracle that features a tunable parameter ranging from 1 to m. This oracle is built to work at a chosen precision level and uses a number of qubits that scales roughly with the problem size and the precision setting. Its overall circuit depth—especially the part requiring resource-intensive non-Clifford gates—is kept modest and, for all settings of the tunable parameter, uses at least about half as many non-Clifford gates as previous designs. In fact, for specific cases like maximum graph cut problems, a smart choice of precision allows us to reduce the circuit depth to grow only logarithmically with the problem size, while keeping qubit connectivity requirements minimal.
Download
Fixed-point Grover Adaptive Search for Quadratic Binary Optimization Problems
Ákos Nagy, Jaime Park, Cindy Zhang, Atithi Acharya, Alex Khan
We explore a variant of Grover’s search tailored for solving Quadratic Unconstrained Binary Optimization (QUBO) problems. For a QUBO problem with n dimensions and m nonzero terms, we design a “marker” oracle that features a tunable parameter ranging from 1 to m. This oracle is built to work at a chosen precision level and uses a number of qubits that scales roughly with the problem size and the precision setting. Its overall circuit depth—especially the part requiring resource-intensive non-Clifford gates—is kept modest and, for all settings of the tunable parameter, uses at least about half as many non-Clifford gates as previous designs. In fact, for specific cases like maximum graph cut problems, a smart choice of precision allows us to reduce the circuit depth to grow only logarithmically with the problem size, while keeping qubit connectivity requirements minimal.
Download
Fixed-point Grover Adaptive Search for Quadratic Binary Optimization Problems
Ákos Nagy, Jaime Park, Cindy Zhang, Atithi Acharya, Alex Khan
We explore a variant of Grover’s search tailored for solving Quadratic Unconstrained Binary Optimization (QUBO) problems. For a QUBO problem with n dimensions and m nonzero terms, we design a “marker” oracle that features a tunable parameter ranging from 1 to m. This oracle is built to work at a chosen precision level and uses a number of qubits that scales roughly with the problem size and the precision setting. Its overall circuit depth—especially the part requiring resource-intensive non-Clifford gates—is kept modest and, for all settings of the tunable parameter, uses at least about half as many non-Clifford gates as previous designs. In fact, for specific cases like maximum graph cut problems, a smart choice of precision allows us to reduce the circuit depth to grow only logarithmically with the problem size, while keeping qubit connectivity requirements minimal.
Download
Anyons in a highly-entangled toric xy model
Milo Moses, Konrad Deka
While ostensibly coined in 1989 by Xiao-Gang Wen, the term "topological order" has been in use since 1972 to describe the behavior of the classical xy model. It has been noted that the xy model does not have Wen's topological order since it is also subject a non-topological U(1) gauge action. We show in a sense this is the only obstruction. That is, if gauge invariance is enforced energetically then the xy model becomes purely topologically ordered. In fact, we show that the quantum xy topological order is an infinite lattice limit of Kitaev's quantum double model applied to the group G=Z.
Download
Anyons in a highly-entangled toric xy model
Milo Moses, Konrad Deka
While ostensibly coined in 1989 by Xiao-Gang Wen, the term "topological order" has been in use since 1972 to describe the behavior of the classical xy model. It has been noted that the xy model does not have Wen's topological order since it is also subject a non-topological U(1) gauge action. We show in a sense this is the only obstruction. That is, if gauge invariance is enforced energetically then the xy model becomes purely topologically ordered. In fact, we show that the quantum xy topological order is an infinite lattice limit of Kitaev's quantum double model applied to the group G=Z.
Download
Anyons in a highly-entangled toric xy model
Milo Moses, Konrad Deka
While ostensibly coined in 1989 by Xiao-Gang Wen, the term "topological order" has been in use since 1972 to describe the behavior of the classical xy model. It has been noted that the xy model does not have Wen's topological order since it is also subject a non-topological U(1) gauge action. We show in a sense this is the only obstruction. That is, if gauge invariance is enforced energetically then the xy model becomes purely topologically ordered. In fact, we show that the quantum xy topological order is an infinite lattice limit of Kitaev's quantum double model applied to the group G=Z.
Download
Creating rotational coherences in molecules aligned along the intermediate moment of inertia axis
Emil J. Zak
We propose and computationally study a method for simultaneously orienting the angular momentum of asymmetric top molecules along: 1) a laboratory-fixed direction; 2) the molecular intermediate moment of inertia axis; 3) the laser field wavevector. For this purpose we utilize a coherent control scheme in which a tailored-pulse optical centrifuge populates rotational states with well defined projections of the total angular momentum onto molecular axes. Appropriately time-shaped optical centrifuge pulses can leave the rotational wavepacket in peculiar rotational coherences which lead to a good degree of 3-dimensional transient alignment, with an arbitrary molecular axis pointing along the laser pulse propagation direction. As an example, we demonstrate how to generate highly resilient rotational quantum states of D2S in which the molecule rotates mainly about its intermediate inertia axis, such that its electric dipole moment is permanently aligned along the propagation direction of the laser pulse. Applications might include accessing less obscured information in various photo-electron imaging experiments.
Download
Creating rotational coherences in molecules aligned along the intermediate moment of inertia axis
Emil J. Zak
We propose and computationally study a method for simultaneously orienting the angular momentum of asymmetric top molecules along: 1) a laboratory-fixed direction; 2) the molecular intermediate moment of inertia axis; 3) the laser field wavevector. For this purpose we utilize a coherent control scheme in which a tailored-pulse optical centrifuge populates rotational states with well defined projections of the total angular momentum onto molecular axes. Appropriately time-shaped optical centrifuge pulses can leave the rotational wavepacket in peculiar rotational coherences which lead to a good degree of 3-dimensional transient alignment, with an arbitrary molecular axis pointing along the laser pulse propagation direction. As an example, we demonstrate how to generate highly resilient rotational quantum states of D2S in which the molecule rotates mainly about its intermediate inertia axis, such that its electric dipole moment is permanently aligned along the propagation direction of the laser pulse. Applications might include accessing less obscured information in various photo-electron imaging experiments.
Download
Creating rotational coherences in molecules aligned along the intermediate moment of inertia axis
Emil J. Zak
We propose and computationally study a method for simultaneously orienting the angular momentum of asymmetric top molecules along: 1) a laboratory-fixed direction; 2) the molecular intermediate moment of inertia axis; 3) the laser field wavevector. For this purpose we utilize a coherent control scheme in which a tailored-pulse optical centrifuge populates rotational states with well defined projections of the total angular momentum onto molecular axes. Appropriately time-shaped optical centrifuge pulses can leave the rotational wavepacket in peculiar rotational coherences which lead to a good degree of 3-dimensional transient alignment, with an arbitrary molecular axis pointing along the laser pulse propagation direction. As an example, we demonstrate how to generate highly resilient rotational quantum states of D2S in which the molecule rotates mainly about its intermediate inertia axis, such that its electric dipole moment is permanently aligned along the propagation direction of the laser pulse. Applications might include accessing less obscured information in various photo-electron imaging experiments.
Download
Benchmarking 16-element quantum search algorithms on superconducting quantum processors
Jan Gwinner, Marcin Briański, Wojciech Burkot, Łukasz Czerwiński, Vladyslav Hlembotskyi
We present experimental results on running 4-qubit unstructured search on IBM quantum processors. Our best attempt attained probability of success around 24.5%. We try several algorithms and use the most recent developments in quantum search to reduce the number of entangling gates that are currently considered the main source of errors in quantum computations. Comparing theoretical expectations of an algorithm performance with the actual data, we explore the hardware limits, showing sharp, phase-transition-like degradation of performance on quantum processors. We conclude that it is extremely important to design hardware-aware algorithms and to include any other low level optimizations on NISQ devices.
Download
Benchmarking 16-element quantum search algorithms on superconducting quantum processors
Jan Gwinner, Marcin Briański, Wojciech Burkot, Łukasz Czerwiński, Vladyslav Hlembotskyi
We present experimental results on running 4-qubit unstructured search on IBM quantum processors. Our best attempt attained probability of success around 24.5%. We try several algorithms and use the most recent developments in quantum search to reduce the number of entangling gates that are currently considered the main source of errors in quantum computations. Comparing theoretical expectations of an algorithm performance with the actual data, we explore the hardware limits, showing sharp, phase-transition-like degradation of performance on quantum processors. We conclude that it is extremely important to design hardware-aware algorithms and to include any other low level optimizations on NISQ devices.
Download
Benchmarking 16-element quantum search algorithms on superconducting quantum processors
Jan Gwinner, Marcin Briański, Wojciech Burkot, Łukasz Czerwiński, Vladyslav Hlembotskyi
We present experimental results on running 4-qubit unstructured search on IBM quantum processors. Our best attempt attained probability of success around 24.5%. We try several algorithms and use the most recent developments in quantum search to reduce the number of entangling gates that are currently considered the main source of errors in quantum computations. Comparing theoretical expectations of an algorithm performance with the actual data, we explore the hardware limits, showing sharp, phase-transition-like degradation of performance on quantum processors. We conclude that it is extremely important to design hardware-aware algorithms and to include any other low level optimizations on NISQ devices.
Download
Efficient unstructured search implementation on current ion-trap quantum processors
Vladyslav Hlembotskyi, Rafał Burczyński, Witold Jarnicki, Adam Szady, Jan Tułowiecki
So far, only the results on 3 qubit spaces (both on superconducting and ion-trap realisations of quantum processors) have beaten the classical unstructured search in the expected number of oracle calls using optimal protocols in both settings. We present experimental results on running unstructured search in spaces defined by 4, 5 and 6 qubits on ion-trapped quantum processor. Our best circuits obtained respectively 66\%, 26\% and 6\% average probability of measuring the marked element. In the case of 4 and 5 qubit spaces we obtained fewer expected number of oracle calls required to find a marked element than any classical approach. Viability of the theoretical result by Grover at these qubit counts is, to authors' knowledge demonstrated experimentally for the first time. Also at 6 qubits, a circuit using a single oracle call returned a measured probability of success exceeding any possible classical approach. These results were achieved using a variety of unstructured search algorithms in conjunction with recent developments in reducing the number of entangling gates. The latter are currently considered to be a dominating source of errors in quantum computations. Some of these improvements have been made possible by using mid-circuit measurements. To our knowledge the latter feature is currently available only on the H0 quantum processor we run on.
Download
Efficient unstructured search implementation on current ion-trap quantum processors
Vladyslav Hlembotskyi, Rafał Burczyński, Witold Jarnicki, Adam Szady, Jan Tułowiecki
So far, only the results on 3 qubit spaces (both on superconducting and ion-trap realisations of quantum processors) have beaten the classical unstructured search in the expected number of oracle calls using optimal protocols in both settings. We present experimental results on running unstructured search in spaces defined by 4, 5 and 6 qubits on ion-trapped quantum processor. Our best circuits obtained respectively 66\%, 26\% and 6\% average probability of measuring the marked element. In the case of 4 and 5 qubit spaces we obtained fewer expected number of oracle calls required to find a marked element than any classical approach. Viability of the theoretical result by Grover at these qubit counts is, to authors' knowledge demonstrated experimentally for the first time. Also at 6 qubits, a circuit using a single oracle call returned a measured probability of success exceeding any possible classical approach. These results were achieved using a variety of unstructured search algorithms in conjunction with recent developments in reducing the number of entangling gates. The latter are currently considered to be a dominating source of errors in quantum computations. Some of these improvements have been made possible by using mid-circuit measurements. To our knowledge the latter feature is currently available only on the H0 quantum processor we run on.
Download
Efficient unstructured search implementation on current ion-trap quantum processors
Vladyslav Hlembotskyi, Rafał Burczyński, Witold Jarnicki, Adam Szady, Jan Tułowiecki
So far, only the results on 3 qubit spaces (both on superconducting and ion-trap realisations of quantum processors) have beaten the classical unstructured search in the expected number of oracle calls using optimal protocols in both settings. We present experimental results on running unstructured search in spaces defined by 4, 5 and 6 qubits on ion-trapped quantum processor. Our best circuits obtained respectively 66\%, 26\% and 6\% average probability of measuring the marked element. In the case of 4 and 5 qubit spaces we obtained fewer expected number of oracle calls required to find a marked element than any classical approach. Viability of the theoretical result by Grover at these qubit counts is, to authors' knowledge demonstrated experimentally for the first time. Also at 6 qubits, a circuit using a single oracle call returned a measured probability of success exceeding any possible classical approach. These results were achieved using a variety of unstructured search algorithms in conjunction with recent developments in reducing the number of entangling gates. The latter are currently considered to be a dominating source of errors in quantum computations. Some of these improvements have been made possible by using mid-circuit measurements. To our knowledge the latter feature is currently available only on the H0 quantum processor we run on.
Download
Introducing Structure to Expedite Quantum Search
Marcin Briański, Jan Gwinner, Vladyslav Hlembotskyi, Witold Jarnicki, Szymon Pliś, Adam Szady
We present a novel quantum algorithm for solving the unstructured search problem with one marked element. Our algorithm allows generating quantum circuits that use asymptotically fewer additional quantum gates than the famous Grover's algorithm and may be successfully executed on NISQ devices. We prove that our algorithm is optimal in the total number of elementary gates up to a multiplicative constant. As many NP-hard problems are not in fact unstructured, we also describe the \emph{partial uncompute} technique which exploits the oracle structure and allows a significant reduction in the number of elementary gates required to find the solution. Combining these results allows us to use asymptotically smaller number of elementary gates than the Grover's algorithm in various applications, keeping the number of queries to the oracle essentially the same. We show how the results can be applied to solve hard combinatorial problems, for example Unique k-SAT. Additionally, we show how to asymptotically reduce the number of elementary gates required to solve the unstructured search problem with multiple marked elements.
Download
Introducing Structure to Expedite Quantum Search
Marcin Briański, Jan Gwinner, Vladyslav Hlembotskyi, Witold Jarnicki, Szymon Pliś, Adam Szady
We present a novel quantum algorithm for solving the unstructured search problem with one marked element. Our algorithm allows generating quantum circuits that use asymptotically fewer additional quantum gates than the famous Grover's algorithm and may be successfully executed on NISQ devices. We prove that our algorithm is optimal in the total number of elementary gates up to a multiplicative constant. As many NP-hard problems are not in fact unstructured, we also describe the \emph{partial uncompute} technique which exploits the oracle structure and allows a significant reduction in the number of elementary gates required to find the solution. Combining these results allows us to use asymptotically smaller number of elementary gates than the Grover's algorithm in various applications, keeping the number of queries to the oracle essentially the same. We show how the results can be applied to solve hard combinatorial problems, for example Unique k-SAT. Additionally, we show how to asymptotically reduce the number of elementary gates required to solve the unstructured search problem with multiple marked elements.
Download
Introducing Structure to Expedite Quantum Search
Marcin Briański, Jan Gwinner, Vladyslav Hlembotskyi, Witold Jarnicki, Szymon Pliś, Adam Szady
We present a novel quantum algorithm for solving the unstructured search problem with one marked element. Our algorithm allows generating quantum circuits that use asymptotically fewer additional quantum gates than the famous Grover's algorithm and may be successfully executed on NISQ devices. We prove that our algorithm is optimal in the total number of elementary gates up to a multiplicative constant. As many NP-hard problems are not in fact unstructured, we also describe the \emph{partial uncompute} technique which exploits the oracle structure and allows a significant reduction in the number of elementary gates required to find the solution. Combining these results allows us to use asymptotically smaller number of elementary gates than the Grover's algorithm in various applications, keeping the number of queries to the oracle essentially the same. We show how the results can be applied to solve hard combinatorial problems, for example Unique k-SAT. Additionally, we show how to asymptotically reduce the number of elementary gates required to solve the unstructured search problem with multiple marked elements.
Download
A short note on graphs with long Thomason's chains
Marcin Briański, Adam Szady
We present a family of 3-connected cubic planar Hamiltonian graphs with an exponential number of steps required by Thomason's algorithm. The base of the exponent is approximately 1.1812..., which exceeds previous results in the area.
Download
A short note on graphs with long Thomason's chains
Marcin Briański, Adam Szady
We present a family of 3-connected cubic planar Hamiltonian graphs with an exponential number of steps required by Thomason's algorithm. The base of the exponent is approximately 1.1812..., which exceeds previous results in the area.
Download
A short note on graphs with long Thomason's chains
Marcin Briański, Adam Szady
We present a family of 3-connected cubic planar Hamiltonian graphs with an exponential number of steps required by Thomason's algorithm. The base of the exponent is approximately 1.1812..., which exceeds previous results in the area.
Download
Our offices
Poland:
Mogilska 43
31-545 Kraków
Canada:
101 College St
Suite H230-1
Toronto
USA:
7757 Baltimore Avenue
Ste 1603
20740 MD College Park
Links
© 2025 BEIT Inc.
Our offices
Poland:
Mogilska 43
31-545 Kraków
Canada:
101 College St
Suite H230-1
Toronto
USA:
7757 Baltimore Avenue
Ste 1603
20740 MD College Park
Links
© 2025 BEIT Inc.
Our offices
Poland:
Mogilska 43
31-545 Kraków
Canada:
101 College St
Suite H230-1
Toronto
USA:
7757 Baltimore Avenue
Ste 1603
20740 MD College Park
Links