Scott Bair is a key voice at Intel Labs, sharing insights into innovative research for inventing tomorrow’s technology.
- Research by beta users of the newly released Intel Quantum SDK 1.0 may contribute to future efforts to speed up complex real-world problem solving.
- Highlighted research efforts by beta users include participants from an Intel Quantum Computing Challenge, and researchers from Deggendorf Institute of Technology and Leidos.
- Intel Quantum SDK will be featured in four research papers presented by Intel Labs at the virtual APS March Meeting on March 21.
Research by beta users of the newly released Intel Quantum Software Development Kit (SDK) version 1.0 may contribute to future efforts to speed up complex problem solving in materials, chemical and drug design, climate modeling, and cryptography.
In early January 2023, beta users participated in an Intel Quantum Computing Challenge held at the 2nd High Performance and Quantum Computing Symposium at Deggendorf Institute of Technology (DIT) in Germany. The submissions explored different quantum use cases, including image denoising and realistic image generation, and solving unstructured search problems. Beyond the challenge, researchers at DIT have used the Intel Quantum SDK to examine a fluid dynamics problem important for aerodynamics and hydrodynamics. In addition, Leidos, an information technology, engineering, and science solutions research company, is exploring computational chemistry and materials modeling as well as theoretical research into thermofield double states.
The preliminary research efforts by beta users are a vital first step in understanding how today’s quantum computing can use a hybrid model consisting of classical and quantum parts, leveraging the unique strengths of each of the two compute models for future large scale quantum computing systems.
“We don't yet have large scale quantum computers, so today we are developing small workloads that we can use to help us design scalable commercial-sized quantum computers of the future,” said Anne Matsuura, director of Quantum Applications and Architecture at Intel Labs. “We develop small quantum algorithms that, in the future, we can scale to simulate materials with unusual electronic properties, such as high temperature superconductors or resistive materials like those used in hard drives. For now, they are useful to help us understand what functionalities need to be provided by the quantum computing system architecture in order to run these algorithms efficiently and accurately on qubits. Working with collaborators in our growing quantum community, we can develop and run these quantum workloads on the small qubit systems we have today.”
The Intel Quantum SDK 1.0 is a full quantum computer in simulation, but it also can interface with Intel’s quantum computing stack. The kit allows developers to program new quantum algorithms for executing qubits in simulation and on real quantum hardware in the future. It includes an intuitive user interface based on C++, a low-level virtual machine (LLVM)-based compiler toolchain with a quantum runtime environment optimized for executing hybrid quantum-classical algorithms and a high-performance Intel Quantum Simulator (IQS) qubit target backend.
Support for the Growing Quantum Community
Intel Quantum SDK 1.0 also includes improvements suggested by beta users, a community of approximately 150 quantum researchers and developers worldwide.
“Intel Quantum SDK is designed specifically for researchers. It’s made by a group of Intel Labs scientists with backgrounds in mathematics, physics, and computer engineering. We speak the same language,” said Rui Li, a professor of thermodynamics and numerical simulation at DIT, and coordinator of the Intel Quantum Computing Challenge. “Quantum SDK allows you to touch the future technology and get to lower levels of abstraction. It’s more open and flexible than other quantum applications.”
The Intel Quantum SDK will be presented at the virtual APS March Meeting on March 20-22, according to Matsuura, who is the program chair for the industrial programming and sessions at APS March Meeting this year. Four Intel Labs papers will be featured on March 21 in Session RR08: V: Quantum Software Stack:
- RR08.00005: Intel Quantum SDK Version 1.0: Extended C++ Compiler, Runtime and Quantum Hardware Simulators for Hybrid Quantum-Classical Applications
- RR08.00006: Efficient Execution of Quantum Algorithms Using the Intel Quantum SDK
- RR08.00007: A Functional Approach to the Modular Construction of Quantum Logic: Part I
- RR08.00008: A Functional Approach to the Modular Construction of Quantum Logic: Part II
Intel Labs also will host an Intel Quantum SDK users meeting in late May.
In addition, Intel Labs also will present papers on quantum computing hardware at the in-person APS March Meeting on March 9 and virtually on March 21:
- TT03.00005: Linear Filtering of Pulses for Cross-Talk Elimination in Frequency-Multiplexed Qubit Control
- W50.001: High-Level Control of Spin Qubits on an Array with 12 Quantum Dots
- T74.007: Si/SiGe Qubit Devices Enabled by Advanced Semiconductor Fabrication
Intel Quantum Computing Challenge Beta User Research
For their work on “Solving 2 by 2 Grid Sudoku Problem Using Grover’s Algorithm,” Tejas Shinde and Yaknan John Gambo, graduate students at DIT, were recognized for their top research submission. Quantum computers can perform a large number of calculations in parallel. This phenomenon has led to the discovery that some algorithms, such as Grover’s algorithm, are faster than classical computers. Sudoku is a combinatorial game based on the Latin square, which has wide applications as an efficient design for controlling multiple sources of variable nuisance simultaneously. Grover’s algorithm, which provides a quadratic speedup, is considered one of the fastest quantum algorithms for searching an unsorted database. It combines superposition and negative amplitudes to solve the unstructured search problem. Grover's algorithm has applications in solving constraint-satisfaction problems, such as Sudoku, type inference, and other logical problem statements. Explorations using Grover’s algorithm could have future implications for cryptography.
For his work on “A Novel Approach to Recreating Markov Chain Using Quantum Circuits for Generative Quantum Machine Learning Applications,” Hossam Ahmed, a researcher at the Leibniz Supercomputing Centre in Munich, was recognized as runner up. Generative machine learning models are trained to generate new data samples similar to the dataset used for training. The model used a sequence generation process, where each step of the process corresponds to a state in the Markov chain and the transition probabilities between states define the probability of generating a particular output at each step. A quantum circuit was used in the Intel Quantum SDK to recreate the Markov chain by including Hadamard gates and connecting different qubits with controlled-NOT (CNOT) gates in order to entangle different qubits. Then a matrix was used to transition between different states. Optimizing the parameters of the quantum circuit helps to improve the performance of the circuit by increasing the accuracy of the output by checking the cost function and reducing the error rate. The reversibility of quantum gates using ancilla bits to store the initial states can yield better results compared to classical models. The research submission showed promising potential for using quantum circuits in various machine learning generative applications such as image denoising and realistic image implementation.
Academic Beta User Research
Beyond coordinating the Intel Quantum Computing Challenge, Li has been collaborating for more than a year on research focused on using quantum computing algorithms for solving flow equations. For “Preliminary Lattice Boltzmann Method Simulation Using Intel Quantum SDK,” his collaborators from DIT include Helena Liebelt, CISO and director of the IT-Centre at the university, and Tejas Shinde, a graduate student. The research was presented at the 21st International Workshop on Advanced Computing and Analysis Techniques in Physics Research (ACAT 2022), which was organized by the European Organization for Nuclear Research (CERN).
Figure 1. Based on Ljubomir Budinski's “Quantum Algorithm for the Advection–Diffusion Equation Simulated with the Lattice Boltzmann Method,” DIT researchers created the above demonstration of the quantum Lattice Boltzmann method.
Transport phenomena such as heat transfer and mass transfer are among today’s most challenging unsolved problems in computational physics due to the inherent nature of fluid complexity. Quantum computing opens a new perspective for numerical simulation including computational fluid dynamics (CFD). Current CFD algorithms based on different macroscopic or microscopic scales need to be translated into a quantum system. For the research, quantum algorithms have been preliminarily implemented for fluid dynamics using the Intel Quantum SDK, and one mesoscopic approach has been applied to solve the Lattice Boltzmann equation. Starting the simplest transport phenomena as a starting point, the preliminary quantum simulation results have been validated with the analytical solution and the classical numerical simulation. This novel approach of using quantum computing to simulate fluid could potentially impact meteorology, materials, energy, and pharmacology.
Industry Beta User Research
For the past six months, researchers at Leidos have collaborated with the Intel Quantum SDK team on preliminary beta research on computational chemistry and materials modeling, and theoretical research into thermofield double states.
“Intel Labs is using a very collaborative approach to develop the Quantum SDK,” said Elizabeth Iwasawa, quantum technology lead and research scientist at the Leidos Innovation Center (LInC). “They want to know what scientists need and what will make the SDK a more useful research tool.”
Leidos electrical engineer Jadyn Bowen and software engineering lead Blake Gage are working on integrating PyTorch, a quantum standard industry machine learning tool, with the Quantum SDK. Using only the Quantum SDK and PyTorch, they ran a classical MNIST dataset and classifier, which is a large database of handwritten digits used for training image processing systems. They created a Quantum SDK analysis environment for editing and reverse-engineering the build script. In the future, they plan on integrating quantum circuits in to the MNIST model and expanding analysis to predict solar data.
Nicholas Stoffle, a senior research scientist at Leidos, is exploring how a one-dimensional quantum Ising model can simulate interactions between neighboring sites in a lattice, and how those changes in magnetization or other effects will propagate throughout the system. This preliminary research may have applications in materials science and monitoring changes in magnetization. For example, this fundamental simulation would allow researchers to model how dramatic temperature changes could affect a new material launched into space. In the future, Stoffle plans on benchmarking the Ising model on other quantum systems and extending to higher dimensions.
Figure 2. Based on Alba Cervera-Lierta's circuit design in the upper left from "Exact Ising Model Simulation on a Quantum Computer," the example circuit shows the work in progress from Leidos on a quantum Ising model simulation.
Leidos scientists also did theoretical research into thermofield double states. Physicist Zachary Guralnik took on the challenge to see if the Intel Quantum SDK could set up a Sachdev-Ye-Kitaev (SYK) model to consider quantum teleportation algorithms, which are believed to be related to the physics of gravity and wormholes.
Accessing the Intel Quantum SDK 1.0
The Intel Quantum SDK 1.0 is available now on the Intel Developer's Cloud.
Debe ser un usuario registrado para añadir un comentario aquí. Si ya está registrado, inicie sesión. Si todavía no está registrado, hágalo e inicie sesión.