AI-Powered Systems Speed Up Quantum Experiments

A graduate student conducting research at MIT has integrated artificial intelligence technology into laboratory software to streamline superconducting qubit experiments and automate laboratory workflows.

Quantum Computers and Experimental Challenges

Quantum computing technology, which offers information processing capacity using quantum mechanics, stands among the most important fields of the future. However, preparing for and conducting experiments for these systems can take months.

Scientists are forced to perform thousands of preliminary measurements during this process, which leads to a significant time cost in research.

The Study Conducted at MIT

Beatriz Yankelevich, a graduate student conducting research within the MIT Engineering Quantum Systems Group, investigated the extent to which artificial intelligence can improve experimental processes.

Within this scope, efforts were aimed at making processes more efficient by utilizing AI-powered tools in studies carried out on superconducting qubits.

Automation in Laboratory Processes

Thanks to connecting the GPT-5.6 Sol model, featuring Codex infrastructure, to laboratory software, the system performed measurements autonomously.

The artificial intelligence also analyzed the obtained results to decide what should be tested in the next phase, providing significant time savings.

Test Results and Research Gains

Tested on an uncalibrated six-qubit chip, the AI system successfully detected transition frequencies when the signals were clear.

The system also managed to determine coherence times by calibrating control pulses, providing great convenience to researchers.

New Focal Points for Researchers

EQuS group researchers have begun regularly utilizing AI agents for routine measurements.

Thanks to this automation, scientists are able to focus on higher-level scientific work such as data interpretation and experimental design.