The IBM Open Science Prize returned to challenge the quantum community once again last November, and the long wait is finally over — the judges have selected the winners.
Founded in 2020, the IBM Quantum Open Science Prize brings the quantum community together to tackle useful problems in the field and present open-source solutions. This year’s theme was quantum state preparation—turning a known quantum state into another known quantum state. The submissions were judged based on performance, scaleability, and most importantly, creativity. The first and second place team winners will receive $30,000 and $20,000, respectively.
Participants were asked to prepare the highly frustrated ground state of a Heisenberg spin-1/2 model on a Kagome lattice using the Variational Quantum Eigensolver (VQE) algorithm on the IBM Quantum Falcon device. The challenge organizers aimed to select a problem that would advance high-fidelity quantum state preparation. Ground states of systems like the one used for this challenge are at the forefront of quantum computing research due to their unique behavior.
The first place prize was awarded to "IBM Open-Science 2022 Qubit Subspace Approach to Kagome" by Tim Weaving, Alexis Ralli, and Vinul Wimalaweera, all from University College London. Their submission effectively utilized advanced methods, specifically qubit tapering and contextual subspace VQE with density matrix renormalization group (DMRG). Qubit tapering decreases the number of required qubits for quantum simulations by taking advantage of symmetries in the system being simulated. Contextual subspace VQE uses both classical and quantum VQE computations to more accurately approximate the ground state energy of a Hamiltonian and additionally reduces the amount of qubits required. This combination of methods successfully reduced the problem size from 12 to 5 qubits.
Then, they ran their 5-qubit experiments in parallel to maximize device throughput. Combining these with quantum error mitigation methods like Zero-Noise Extrapolation and Readout Error Mitigation, they achieved a significant improvement in fidelity.
The runner-up prize was awarded to "Ground state of S-1/2 IBM KHA" by Pratyay Ghosh, Alexander Fritzsche, Alexander Stegmaier, Richard Strunck, and Jannis Seufert, all from Julius-Maximilians-Universität Würzburg. Their submission thoroughly exploited the symmetries of the problem Hamiltonian, reducing their ansatz circuit to only 4 parameters. Combining this with a novel error mitigation scheme helped to improve the fidelity when run on the quantum processor.
Both winning teams broke the problem down into smaller pieces and effectively used advanced error mitigation techniques. This competition also fueled the rest of the community to learn more about these complex topics — we received more than 130 submissions.
This is the third of three successful Open Science Prizes, each one providing important insights that have helped the quantum community at large. In 2020, the challenge was to either to reduce SWAP gate errors or to improve the fidelity of graph state preparation on an IBM Quantum device. In 2021, the challenge was to simulate a Heisenberg model Hamiltonian for a three-particle system on IBM Quantum hardware using Qiskit Pulse or Qiskit defaults. 2022’s challenge was no exception, making participants come up with creative ways to implement VQE on the highest qubit IBM device used so far for the Open Science Prize. They demonstrated that the quantum community will continue to advance the field and help us on the path to useful quantum computing.
The search for utility in quantum-centric supercomputing is far from over. We need researchers and developers to continue to address the largest problems in the field. Missed this year’s competition? Keep an eye out for our next Open Science Prize and continue contributing to our growing quantum community.
