Mid-circuit measurement with QubiCML with a readout time of 500 ns.
As quantum information processors continue to expand in both quantum bit (qubit) count and functionality, the control and measurement systems for qubits pose scalability challenges. Unrestricted access to the entire control stack becomes imperative for comprehensive system-level optimization.
BACI researchers have developed a patent-pending QubiCML for the QubiC system that combines traditional electronic control circuits, including field-programmable gate arrays (FPGAs), with machine learning algorithms. QubiCML enables accurate, real-time discrimination of the state of superconducting qubits during intermediate stages of a quantum circuit. It could help bring quantum computing closer to reality and provide ultra-high-precision control technology for particle accelerators and laser systems.
By performing mid-circuit quantum state discrimination—where a qubit’s state is measured at predetermined intervals during a quantum circuit’s operation—QubiCML could reduce the time qubits need to maintain coherence, enabling real-time error detection and correction. It could also help replace some quantum operations with classical logic, simplifying quantum circuits.
The QubiC system continues to evolve to serve the needs of quantum processing.
In tests, QubiCML took only 54 nanoseconds to perform each qubit inference (state measurement) on three superconducting qubits, with an average accuracy of 98.46%. Additionally, the readout time was just 500 nanoseconds, which is significantly faster than the qubit’s coherence time and is considered cutting-edge in the quantum computing community.
QubiCML builds on and extends the capabilities of QubiC, an open-source FPGA-based control system created by Berkeley Lab. QubiC includes digital-to-analog converters to generate radio-frequency pulses to control qubits, analog-to-digital converters to measure qubit responses, and digital signal-processing units on an FPGA to analyze the data.
QubiC is supported by the Advanced Quantum Testbed and the Quantum Systems Accelerator, both hosted at Berkeley Lab and funded by the U.S. Department of Energy (DOE)’s Office of Science, Office of Advanced Scientific Computing Research, the Quantum Information Science Enabled Discovery program funded by the DOE’s Office of Science, High Energy Physics, and the DOE’s Office of Science Laboratory Directed Research and Development program.
