As a fundamental tool in AI, convex optimization has been a significant research field for many years, and the same goes for its online version. Recently, general convex optimization problem has been accelerated with the help of quantum computing, and the technique of online convex optimization has been used for accelerating the online quantum state learning problem, thus we want to study whether online convex optimization (OCO) model can also be benefited from quantum computing. In this paper, we consider the OCO model, which can be described as a $T$ round iterative game between the player and the adversary. A key factor for measuring the performance of an OCO algorithm ${\cal A}$ is the regret denoted by $\text{regret}_{T}(\mathcal{A})$ , and it is said to perform well if its regret is sublinear as a function of $T$ . Another factor is the computational cost (e.g., query complexity) of the algorithm. We give a quantum algorithm for the online convex optimization model with only zeroth-order oracle available, which can achieve $O(\sqrt{T})$ and $O(\log{T})$ regret for general convex loss functions and $\alpha$ -strong loss functions respectively, where only $O(1)$ queries are needed in each round. Our results show that the zeroth-order quantum oracle is as powerful as the classical first-order oracle, and show potential advantages of quantum computing over classical computing in the OCO model where only zeroth-order oracle available.As a fundamental tool in AI, convex optimization has been a significant research field for many years, and the same goes for its online version. Recently, general convex optimization problem has been accelerated with the help of quantum computing, and the technique of online convex optimization has been used for accelerating the online quantum state learning problem, thus we want to study whether online convex optimization (OCO) model can also be benefited from quantum computing. In this paper, we consider the OCO model, which can be described as a $T$ round iterative game between the player and the adversary. A key factor for measuring the performance of an OCO algorithm ${\cal A}$ is the regret denoted by $\text{regret}_{T}(\mathcal{A})$, and it is said to perform well if its regret is sublinear as a function of $T$. Another factor is the computational cost (e.g., query complexity) of the algorithm. We give a quantum algorithm for the online convex optimization model with only zeroth-order oracle available, which can achieve $O(\sqrt{T})$ and $O(\log{T})$ regret for general convex loss functions and $\alpha$-strong loss functions respectively, where only $O(1)$ queries are needed in each round. Our results show that the zeroth-order quantum oracle is as powerful as the classical first-order oracle, and show potential advantages of quantum computing over classical computing in the OCO model where only zeroth-order oracle available.

Jul 29, 20201 citation

Electronic Quantum Computing

David Peter Divincenzo

Quantum Electronics and Laser Science ConferenceMay 21, 20060 citations

Quantum Online Algorithms with Respect to Space and Advice Complexity

Kamil Ravilevich Khadiev, Aliya Khadieva, Ilnaz Mannapov

Online computation is a well-known area of computer science. We introduce quantum online algorithms and investigate them with respect to a competitive ratio in two points of view: space complexity and advice complexity. We start with exploring a model with restricted memory and show that quantum online algorithms can be better than classical ones (deterministic or randomized) for sublogarithmic space (memory), and they can be better than deterministic online algorithms without restriction for memory. Additionally, we consider the polylogarithmic space case and show that in this case, quantum online algorithms can be better than deterministic ones as well. Another point of view to the online algorithms model is advice complexity. So, we introduce quantum online algorithms with a quantum channel with an adviser. Firstly, we show that quantum algorithms have at least the same computational power as classical ones have. We give some examples of quantum online algorithms with advice. Secondly, we show that if we allow to use shared entangled qubits (EPR-pairs), then a quantum online algorithm can use half as many advise qubits compared to a classical one. We apply this approach to the well-known Paging Problem.Online computation is a well-known area of computer science. We introduce quantum online algorithms and investigate them with respect to a competitive ratio in two points of view: space complexity and advice complexity. We start with exploring a model with restricted memory and show that quantum online algorithms can be better than classical ones (deterministic or randomized) for sublogarithmic space (memory), and they can be better than deterministic online algorithms without restriction for memory. Additionally, we consider the polylogarithmic space case and show that in this case, quantum online algorithms can be better than deterministic ones as well. Another point of view to the online algorithms model is advice complexity. So, we introduce quantum online algorithms with a quantum channel with an adviser. Firstly, we show that quantum algorithms have at least the same computational power as classical ones have. We give some examples of quantum online algorithms with advice. Secondly, we show that if we allow to use shared entangled qubits (EPR-pairs), then a quantum online algorithm can use half as many advise qubits compared to a classical one. We apply this approach to the well-known Paging Problem.

Lobachevskii Journal of MathematicsNov 1, 201832 citations

Locality in distributed quantum and online graph algorithms

Henrik Lievonen

In this thesis, I study how quantum computation and quantum communication can help large networks of computers to solve computational problems faster. I also connect the notion of locality between various models, including distributed quantum algorithms and online graph algorithms, and show that locality in online graph algorithms can be used to provide bounds for distributed quantum advantage. Towards proving distributed quantum advantage, I present the first family of LCL problems which shows a quantum advantage for distributed graph algorithms. Here LCLs, or Locally Checkable Labeling problems, are a family of graph problems described by a finite set of local constraints, and I mainly focus on them in this thesis. The models of distributed computing which this result separates are the classical LOCAL model and the quantum-LOCAL model. In both of these models, a network of computers exchanges messages in synchronous rounds, and the running time of the algorithm directly corresponds with locality, that is, how far the computers see in the network. In this thesis, I also introduce new models of computing that capture locality in stronger settings. Most notably, I introduce the online-LOCAL model which captures the locality of online graph algorithms, and show that it is, asymptotically, roughly as strong as the classical LOCAL model for solving LCL problems on rooted regular trees. I provide an online-LOCAL algorithm for finding a 3-coloring of a bipartite graph with lower locality than is possible in the classical LOCAL model, and show that the algorithm is tight. To draw things together, I show that quantum-LOCAL and online-LOCAL are indeed related to each other: any quantum-LOCAL algorithm can be simulated in the randomized online-LOCAL model. The converse does not hold in general, but an online-LOCAL algorithm solving an LCL problem in rooted, possibly non-regular, trees can be simulated in the classical LOCAL model—and thus in the quantum-LOCAL model—with an exponential cost in locality, putting indirectly a bound on the maximum possible attainable quantum advantage. Finally, I consider the sinkless orientation problem, which holds a central role in our understanding of the locality of LCL problems in the classical LOCAL model, and provide a simpler and more direct proof for its hardness using the supported LOCAL model. I also provide a simpler algorithm solving sinkless orientation in the SLOCAL model, which is a sequential version of the LOCAL model.In this thesis, I study how quantum computation and quantum communication can help large networks of computers to solve computational problems faster. I also connect the notion of locality between various models, including distributed quantum algorithms and online graph algorithms, and show that locality in online graph algorithms can be used to provide bounds for distributed quantum advantage. Towards proving distributed quantum advantage, I present the first family of LCL problems which shows a quantum advantage for distributed graph algorithms. Here LCLs, or Locally Checkable Labeling problems, are a family of graph problems described by a finite set of local constraints, and I mainly focus on them in this thesis. The models of distributed computing which this result separates are the classical LOCAL model and the quantum-LOCAL model. In both of these models, a network of computers exchanges messages in synchronous rounds, and the running time of the algorithm directly corresponds with locality, that is, how far the computers see in the network. In this thesis, I also introduce new models of computing that capture locality in stronger settings. Most notably, I introduce the online-LOCAL model which captures the locality of online graph algorithms, and show that it is, asymptotically, roughly as strong as the classical LOCAL model for solving LCL problems on rooted regular trees. I provide an online-LOCAL algorithm for finding a 3-coloring of a bipartite graph with lower locality than is possible in the classical LOCAL model, and show that the algorithm is tight. To draw things together, I show that quantum-LOCAL and online-LOCAL are indeed related to each other: any quantum-LOCAL algorithm can be simulated in the randomized online-LOCAL model. The converse does not hold in general, but an online-LOCAL algorithm solving an LCL problem in rooted, possibly non-regular, trees can be simulated in the classical LOCAL model—and thus in the quantum-LOCAL model—with an exponential cost in locality, putting indirectly a bound on the maximum possible attainable quantum advantage. Finally, I consider the sinkless orientation problem, which holds a central role in our understanding of the locality of LCL problems in the classical LOCAL model, and provide a simpler and more direct proof for its hardness using the supported LOCAL model. I also provide a simpler algorithm solving sinkless orientation in the SLOCAL model, which is a sequential version of the LOCAL model.

Jan 1, 20260 citations

Practical Quantum Computing

Patrick M. Lenahan

Elsevier eBooksJan 1, 20031 citation

T

round iterative game between the player and the adversary. A key factor for measuring the performance of an OCO algorithm

{\cal A}

is the regret denoted by

\text{regret}_{T}(\mathcal{A})

, and it is said to perform well if its regret is sublinear as a function of

T

. Another factor is the computational cost (e.g., query complexity) of the algorithm. We give a quantum algorithm for the online convex optimization model with only zeroth-order oracle available, which can achieve

O(\sqrt{T})

and

O(\log{T})

regret for general convex loss functions and

\alpha

-strong loss functions respectively, where only

O(1)

queries are needed in each round. Our results show that the zeroth-order quantum oracle is as powerful as the classical first-order oracle, and show potential advantages of quantum computing over classical computing in the OCO model where only zeroth-order oracle available.