By the Qiskit Applications Team
Natural Sciences and Quantum Simulation
Richard Feynman envisioned quantum devices as a new computational paradigm for the simulation of Nature at its fundamental (quantum) level. Today, we are announcing Qiskit Nature, a unique platform to bridge the gap between natural sciences and quantum simulations. The modularity of Qiskit Nature will allow researchers in different areas of natural sciences (including physics, chemistry, material science and biology) to model and solve domain-specific problems using quantum simulations as well as direct experiments on quantum computers via the IBM Quantum Cloud systems. In addition, the user interface will enable researchers to interact with Qiskit Nature at different levels, ranging from software development (allowing direct programming of new functionalities) to solution-driven applications (requiring no specific quantum computing or quantum hardware knowledge)
Qiskit Nature
Qiskit Nature is designed for generic users — ranging from quantum computing experts to more applied, domain expert researchers — to investigate and simulate problems in natural sciences using quantum algorithms. Qiskit Nature’s modularity and its wide composition of different modules ensures its extensive capabilities to address complex problems in physics, chemistry and biology. In upcoming releases we will fully exploit this potential by introducing more applications related to natural sciences.
As a first step toward a modular Qiskit Nature, we’re re-structuring Qiskit Chemistry to fit into this broader framework. While today’s functionality is similar, you’ll notice that way that you define a problem of interest is different. We encourage you to read the documentation to familiarize yourself with the new structures.
Natural Sciences Problems
We designed Qiskit Nature to guarantee modularity and extensibility, while keeping a user-friendly interface that enables its direct application to real world problems. New in Qiskit Nature is the notion of a ‘Problem’, which is a class that identifies particular, domain-specific quantum computing solutions that leverage general quantum algorithms supplied in Qiskit 0.25 — “electronic structure problem,” for example. The Problem-based structure enables a straightforward extension of Qiskit Nature toward other problem classes in condensed matter, lattice field theories, and biology.
This first release focuses on quantum chemistry solutions for electronic structure and vibrational structure problems. The built-in mapping of the initial fermionic and bosonic second quantized Hamiltonian into qubit operators leverages different levels of optimization and delivers corresponding qubit operators without the need of specific quantum computing competence.
Encoding Your Problem
As part of this release, Qiskit Nature will allow us to use the ubiquitous second quantization formalism to solve a variety of problems, ranging from fundamental particle physics to solid state physics, quantum chemistry and statistical mechanics. Qiskit Nature helps you re-frame your problem using operators that fulfill fermionic, bosonic or spin statistics. Combinations of degrees of freedom subject to different commutation rules are also supported. This functionality will enable the implementation of arbitrary Hamiltonians — or, in short, Qiskit Nature is a stepping stone toward developing new applications in the natural sciences domain.
Using Chemistry Drivers With Qiskit
Tackling quantum chemistry problems with Qiskit Nature requires the use of classical code, such as the PySCF driver. Qiskit Nature makes it easy to import and use these drivers, and then act on the output of the drivers using Transformers. Transformers help select or modify only the quantities from the Driver that you want to use, and then return an updated input to use for your quantum chemistry problem. Essentially, these transformers are ensuring that your program is only using the pieces of the classical data necessary in order to maximize its efficiency. Examples of transformations applied to electronic structure problems are the active space and freeze core reduction schemes. All future operations affecting the generation of second quantized operators will become part of this interface.
Mapping Problems to Qubits
Once we have defined the problem, Qiskit Nature also helps map it onto the quantum computer. Second quantized operators are mapped onto qubit operators that can be directly used in quantum computations, without losing the information related to the nature of the problem. Different mappings are required for each different problem class and are available in Qiskit Nature via the qubit converter. In the case of fermionic problems, three different mapping schemes are provided: Jordan Wigner, Bravyi Kitaev, and Parity. For the spin operators we support the linear mapping while for vibrational operators the direct mapping is provided. The qubit converter can also perform qubit reduction operations exploiting symmetries in the Hamiltonian (expressed in qubit form) or other intrinsic properties of the qubit operator.
How to get started
We encourage you to begin familiarizing yourself with the Problem class, and to explore the tools introduced today by following this link — https://qisk.it/2PDD39L — and reading the newly released tutorial notebooks made available here: https://qisk.it/3fMMMVV. In addition, you can check out the Qiskit Slack workspace to connect with other Qiskit users and contributors. If you’ve gotten a little lost in the Quantum Chemistry jargon, don’t worry; we’ll soon have other Qiskit Nature applications that will fit your specific domain’s needs.
A special thanks to the core contributors (in alphabetical order):
Panagiotis Barkoutsos, Anton Dekusar, Bryce Fuller, Julien Gacon, Ikko Hamamura, Takashi Imamichi, John Lapeyre, Dariusz Lasecki, Manoel Marques, Simon Mathis, Atsushi Matsuo, Giulia Mazzola, Pauline Ollitrault, Max Rossmannek, Igor Sokolov, Ivano Tavernelli, Stefan Woerner, Steve Wood
