The Qiskit Primitives is the most recent release of Qiskit; this is considered a new approach to communicating with quantum computers. The latest release includes a completely redesigned algorithms module, which offers state-of-the-art quantum algorithms for the open-source community, all driven by the newly introduced primitives.
All changes in Qiskit’s algorithms module?
The minimum eigensolvers module handles the interfaces and quantum algorithms that compute a Hamiltonian’s ground state. The update corrects a minor naming error in the old minimum eigensolvers and includes the new primitive-based algorithms. It contains well-known optimization techniques such as VQE and QAOA, which now support the Qiskit Primitives.
It’s worth mentioning that starting with this release; there have been two versions of VQE available: SamplingVQE, which utilizes the sampler primitive and is optimized for diagonal Hamiltonians, and VQE, which uses the estimator primitive. The updated VQE algorithm now employs an estimator rather than a quantum instance and the SparsePauliOp rather than the PauliSumOp.
AdaptVQE from Qiskit Nature was merged into Qiskit v0.39 as a new method introduced in this release and generalized beyond the chemistry applications. It computes a system’s ground state energy by constructing a compact ansatz from a set of evolution operators.
The eigensolvers module contains interfaces and techniques for calculating operator eigenvalues. This version, once again, corrects a minor nomenclature discrepancy in the old eigensolvers module and includes the new primitive-based algorithms. For example, they can be used to determine the energy spectra of quantum systems. For example, they can be used to determine the energy spectra of quantum systems. Qiskit now provides a reference classical implementation based on NumPy and a quantum approach known as Variational Quantum Deflation (VQD). VQD computes excited state energies of Hamiltonians and is a new addition to this version’s module.
The time evolvers module now contains primitive-enabled time evolution algorithms and interfaces. This includes the Trotterization-based Quantum Real-Time Evolution algorithm and the Projected Variational Quantum Dynamics (pVQD) algorithm, which is new in this release.
Our developers exploited the release of Qiskit primitive programs to migrate features from the gradient framework, which was already accessible in qiskit.opflow, to a new module with expanded capabilities. They also created another valid subroutine for developing quantum algorithms: calculating quantum state fidelities. They make it possible for every algorithm developer to reap the benefits of Qiskit Primitives in their work.
Developers exploited the release of Qiskit primitive programs to migrate features from the gradient framework, which was already accessible in qiskit.opflow, to a new module with expanded capabilities. They also created another valid subroutine for developing quantum algorithms: calculating quantum state fidelities. They make it possible for every algorithm developer to reap the benefits of Qiskit Primitives in their work.
Many quantum algorithms necessitate the computation of gradients, such as sampling probabilities or expectation values. This led to the development of a new algorithm module called gradients, making it easier to use Qiskit Primitives. Estimator gradients allow for the computation of gradients of observables, while sampler gradients allow for the calculation of gradients of sampled probabilities. The supported gradient methods are finite difference, a linear combination of unitaries, parameters shift, and SPSA. Also, recently, the new class allows for Quantum Fisher Information (QFI) calculation.
In this version, the developers also included a procedure for using Qiskit Primitives. The compute-computer fidelity calculation method is currently available in the state fidelities package.
Read more about it here.